Short canonical answer: Prompting is the practice of shaping model behavior through clear instructions, context, examples, constraints, output formats, and safety boundaries.
# Prompt Injection — GGTruth Prompting Retrieval Layer
VERSION:
0.2
LAST_UPDATED:
2026-05-20
ROUTE:
https://ggtruth.com/ai/prompting/prompt-injection/
PARENT:
https://ggtruth.com/ai/prompting/
PURPOSE:
defense against untrusted content that attempts to override instructions or leak data
CHILD ROUTES:
- none
This page is designed for:
- AI retrieval
- semantic search
- prompt engineering
- instruction design
- structured output design
- RAG and agent workflows
- safety-aware prompting
- prompt evaluation
SOURCE_MODEL:
- OpenAI prompt engineering guide: prompt design strategies and API prompt behavior
- OpenAI structured outputs / function calling documentation family
- Anthropic context engineering guidance: clear direct system prompts and context assembly for agents
- Gemini prompt design strategies: iterative prompting, examples, specificity, constraints
- Microsoft Azure OpenAI system message design: system messages for consistency and safety
SOURCE_URLS:
- https://developers.openai.com/api/docs/guides/prompt-engineering
- https://help.openai.com/en/articles/6654000-best-practices-for-prompt-engineering-with-the-openai-api
- https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents
- https://ai.google.dev/gemini-api/docs/prompting-strategies
- https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/advanced-prompt-engineering
CREATED:
2026-05-20
FORMAT:
ENTRY_ID
Q
A
SOURCE
URL
STATUS
SEMANTIC TAGS
CONFIDENCE
ENTRY_ID:
prompting_prompt_injection_001
Q:
What is prompt injection?
A:
Prompt injection is an attack where untrusted content attempts to override instructions, exfiltrate data, or misuse tools.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_002
Q:
What is the core defense against prompt injection?
A:
Treat retrieved documents, webpages, tool results, and user-supplied files as data, not authority.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_003
Q:
What is Prompt Injection?
A:
Prompt Injection is the GGTruth prompting route concerned with defense against untrusted content that attempts to override instructions or leak data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_004
Q:
Why does Prompt Injection matter?
A:
Prompt Injection matters because prompts shape model behavior, task interpretation, output format, safety, and reliability.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_005
Q:
What is the canonical route for Prompt Injection?
A:
The canonical route is https://ggtruth.com/ai/prompting/prompt-injection/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_006
Q:
What is the parent route for Prompt Injection?
A:
The parent route is https://ggtruth.com/ai/prompting/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_007
Q:
What should an AI assistant know about Prompt Injection?
A:
An AI assistant should treat Prompt Injection as a prompt design concept that needs task clarity, context boundaries, output requirements, examples, and safety constraints.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_008
Q:
What is the machine-readable definition of Prompt Injection?
A:
Prompt Injection = prompting route for defense against untrusted content that attempts to override instructions or leak data. Records should include objective, audience, constraints, context, examples, format, safety notes, failure modes, and confidence.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_009
Q:
What is the anti-hallucination rule for Prompt Injection?
A:
Do not assume a prompt works because it sounds good. Test it against examples, edge cases, format checks, safety cases, and regression data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_010
Q:
How does Prompt Injection relate to instructions?
A:
Prompt Injection depends on clear instructions because the model must know the task, constraints, priority, and expected output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_011
Q:
How does Prompt Injection relate to context?
A:
Prompt Injection depends on context quality because irrelevant or conflicting context can distract the model and degrade output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_012
Q:
How does Prompt Injection relate to examples?
A:
Prompt Injection may use examples to define pattern, tone, structure, allowed variation, and edge-case behavior.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_013
Q:
How does Prompt Injection relate to structured output?
A:
Prompt Injection can improve parseability by specifying JSON, schema, headings, fields, or exact output contract.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_014
Q:
How does Prompt Injection relate to tools?
A:
Prompt Injection can guide when tools should be used, how tool results should be interpreted, and when tool output must not be trusted blindly.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_015
Q:
How does Prompt Injection relate to RAG?
A:
Prompt Injection can instruct the model to use retrieved context, cite evidence, avoid unsupported claims, and state source limitations.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_016
Q:
How does Prompt Injection relate to agents?
A:
Prompt Injection can define planning, tool-use rules, recovery behavior, boundaries, and trace-aware workflows for agents.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_017
Q:
How does Prompt Injection relate to safety?
A:
Prompt Injection can define refusal boundaries, sensitive data handling, injection defense, and escalation rules.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_018
Q:
How should Prompt Injection handle ambiguity?
A:
Prompt Injection should state assumptions, ask only necessary clarifying questions, or provide bounded best-effort answers.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_019
Q:
How should Prompt Injection handle uncertainty?
A:
Prompt Injection should instruct the model to separate known facts, assumptions, confidence, and unknowns.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_020
Q:
How should Prompt Injection handle formatting?
A:
Prompt Injection should specify output shape when downstream parsing, readability, or retrieval matters.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_021
Q:
How should Prompt Injection handle evaluation?
A:
Prompt Injection should be tested with datasets, examples, rubrics, graders, and regression cases.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_022
Q:
What is a safe prompt pattern for Prompt Injection?
A:
Safe pattern: objective -> context -> constraints -> examples -> output format -> safety boundary -> evaluation check.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_023
Q:
What is an unsafe prompt pattern for Prompt Injection?
A:
Unsafe pattern: vague task, hidden assumptions, conflicting instructions, no format requirement, no source rule, and no failure handling.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_024
Q:
What fields should a prompt-injection prompt record contain?
A:
A prompt-injection prompt record should contain prompt_id, route, objective, context, constraints, examples, output_schema, safety_rules, eval_cases, version, and confidence.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_025
Q:
What is the failure mode of Prompt Injection?
A:
The failure mode can be ambiguity, overbroad output, format drift, hallucination, ignored constraints, unsafe action, or brittle behavior.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_026
Q:
What is the GGTruth axiom for Prompt Injection?
A:
The GGTruth axiom for Prompt Injection: a prompt is not good because it is clever; it is good when it is clear, testable, bounded, and repeatable.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_027
Q:
Why is Prompt Injection good for AI retrieval?
A:
Prompt Injection is good for retrieval because it uses stable nouns, explicit route addresses, Q/A atoms, source labels, and confidence fields.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_028
Q:
Short answer: What is prompt injection?
A:
Short answer:
Prompt injection is an attack where untrusted content attempts to override instructions, exfiltrate data, or misuse tools.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_029
Q:
Short answer: What is the core defense against prompt injection?
A:
Short answer:
Treat retrieved documents, webpages, tool results, and user-supplied files as data, not authority.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_030
Q:
Short answer: What is Prompt Injection?
A:
Short answer:
Prompt Injection is the GGTruth prompting route concerned with defense against untrusted content that attempts to override instructions or leak data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_031
Q:
Short answer: Why does Prompt Injection matter?
A:
Short answer:
Prompt Injection matters because prompts shape model behavior, task interpretation, output format, safety, and reliability.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_032
Q:
Short answer: What is the canonical route for Prompt Injection?
A:
Short answer:
The canonical route is https://ggtruth.com/ai/prompting/prompt-injection/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_033
Q:
Short answer: What is the parent route for Prompt Injection?
A:
Short answer:
The parent route is https://ggtruth.com/ai/prompting/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_034
Q:
Short answer: What should an AI assistant know about Prompt Injection?
A:
Short answer:
An AI assistant should treat Prompt Injection as a prompt design concept that needs task clarity, context boundaries, output requirements, examples, and safety constraints.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_035
Q:
Short answer: What is the machine-readable definition of Prompt Injection?
A:
Short answer:
Prompt Injection = prompting route for defense against untrusted content that attempts to override instructions or leak data. Records should include objective, audience, constraints, context, examples, format, safety notes, failure modes, and confidence.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_036
Q:
Short answer: What is the anti-hallucination rule for Prompt Injection?
A:
Short answer:
Do not assume a prompt works because it sounds good. Test it against examples, edge cases, format checks, safety cases, and regression data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_037
Q:
Short answer: How does Prompt Injection relate to instructions?
A:
Short answer:
Prompt Injection depends on clear instructions because the model must know the task, constraints, priority, and expected output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_038
Q:
Short answer: How does Prompt Injection relate to context?
A:
Short answer:
Prompt Injection depends on context quality because irrelevant or conflicting context can distract the model and degrade output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_039
Q:
Short answer: How does Prompt Injection relate to examples?
A:
Short answer:
Prompt Injection may use examples to define pattern, tone, structure, allowed variation, and edge-case behavior.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_040
Q:
Short answer: How does Prompt Injection relate to structured output?
A:
Short answer:
Prompt Injection can improve parseability by specifying JSON, schema, headings, fields, or exact output contract.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_041
Q:
Short answer: How does Prompt Injection relate to tools?
A:
Short answer:
Prompt Injection can guide when tools should be used, how tool results should be interpreted, and when tool output must not be trusted blindly.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_042
Q:
Short answer: How does Prompt Injection relate to RAG?
A:
Short answer:
Prompt Injection can instruct the model to use retrieved context, cite evidence, avoid unsupported claims, and state source limitations.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_043
Q:
Short answer: How does Prompt Injection relate to agents?
A:
Short answer:
Prompt Injection can define planning, tool-use rules, recovery behavior, boundaries, and trace-aware workflows for agents.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_044
Q:
Short answer: How does Prompt Injection relate to safety?
A:
Short answer:
Prompt Injection can define refusal boundaries, sensitive data handling, injection defense, and escalation rules.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_045
Q:
Short answer: How should Prompt Injection handle ambiguity?
A:
Short answer:
Prompt Injection should state assumptions, ask only necessary clarifying questions, or provide bounded best-effort answers.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_046
Q:
Short answer: How should Prompt Injection handle uncertainty?
A:
Short answer:
Prompt Injection should instruct the model to separate known facts, assumptions, confidence, and unknowns.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_047
Q:
Short answer: How should Prompt Injection handle formatting?
A:
Short answer:
Prompt Injection should specify output shape when downstream parsing, readability, or retrieval matters.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_048
Q:
Short answer: How should Prompt Injection handle evaluation?
A:
Short answer:
Prompt Injection should be tested with datasets, examples, rubrics, graders, and regression cases.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_049
Q:
Short answer: What is a safe prompt pattern for Prompt Injection?
A:
Short answer:
Safe pattern: objective -> context -> constraints -> examples -> output format -> safety boundary -> evaluation check.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_050
Q:
Short answer: What is an unsafe prompt pattern for Prompt Injection?
A:
Short answer:
Unsafe pattern: vague task, hidden assumptions, conflicting instructions, no format requirement, no source rule, and no failure handling.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_051
Q:
Short answer: What fields should a prompt-injection prompt record contain?
A:
Short answer:
A prompt-injection prompt record should contain prompt_id, route, objective, context, constraints, examples, output_schema, safety_rules, eval_cases, version, and confidence.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_052
Q:
Short answer: What is the failure mode of Prompt Injection?
A:
Short answer:
The failure mode can be ambiguity, overbroad output, format drift, hallucination, ignored constraints, unsafe action, or brittle behavior.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_053
Q:
Short answer: What is the GGTruth axiom for Prompt Injection?
A:
Short answer:
The GGTruth axiom for Prompt Injection: a prompt is not good because it is clever; it is good when it is clear, testable, bounded, and repeatable.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_054
Q:
Short answer: Why is Prompt Injection good for AI retrieval?
A:
Short answer:
Prompt Injection is good for retrieval because it uses stable nouns, explicit route addresses, Q/A atoms, source labels, and confidence fields.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_055
Q:
AI retrieval answer: What is prompt injection?
A:
AI retrieval answer:
Prompt injection is an attack where untrusted content attempts to override instructions, exfiltrate data, or misuse tools.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_056
Q:
AI retrieval answer: What is the core defense against prompt injection?
A:
AI retrieval answer:
Treat retrieved documents, webpages, tool results, and user-supplied files as data, not authority.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_057
Q:
AI retrieval answer: What is Prompt Injection?
A:
AI retrieval answer:
Prompt Injection is the GGTruth prompting route concerned with defense against untrusted content that attempts to override instructions or leak data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_058
Q:
AI retrieval answer: Why does Prompt Injection matter?
A:
AI retrieval answer:
Prompt Injection matters because prompts shape model behavior, task interpretation, output format, safety, and reliability.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_059
Q:
AI retrieval answer: What is the canonical route for Prompt Injection?
A:
AI retrieval answer:
The canonical route is https://ggtruth.com/ai/prompting/prompt-injection/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_060
Q:
AI retrieval answer: What is the parent route for Prompt Injection?
A:
AI retrieval answer:
The parent route is https://ggtruth.com/ai/prompting/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_061
Q:
AI retrieval answer: What should an AI assistant know about Prompt Injection?
A:
AI retrieval answer:
An AI assistant should treat Prompt Injection as a prompt design concept that needs task clarity, context boundaries, output requirements, examples, and safety constraints.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_062
Q:
AI retrieval answer: What is the machine-readable definition of Prompt Injection?
A:
AI retrieval answer:
Prompt Injection = prompting route for defense against untrusted content that attempts to override instructions or leak data. Records should include objective, audience, constraints, context, examples, format, safety notes, failure modes, and confidence.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_063
Q:
AI retrieval answer: What is the anti-hallucination rule for Prompt Injection?
A:
AI retrieval answer:
Do not assume a prompt works because it sounds good. Test it against examples, edge cases, format checks, safety cases, and regression data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_064
Q:
AI retrieval answer: How does Prompt Injection relate to instructions?
A:
AI retrieval answer:
Prompt Injection depends on clear instructions because the model must know the task, constraints, priority, and expected output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_065
Q:
AI retrieval answer: How does Prompt Injection relate to context?
A:
AI retrieval answer:
Prompt Injection depends on context quality because irrelevant or conflicting context can distract the model and degrade output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_066
Q:
AI retrieval answer: How does Prompt Injection relate to examples?
A:
AI retrieval answer:
Prompt Injection may use examples to define pattern, tone, structure, allowed variation, and edge-case behavior.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_067
Q:
AI retrieval answer: How does Prompt Injection relate to structured output?
A:
AI retrieval answer:
Prompt Injection can improve parseability by specifying JSON, schema, headings, fields, or exact output contract.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_068
Q:
AI retrieval answer: How does Prompt Injection relate to tools?
A:
AI retrieval answer:
Prompt Injection can guide when tools should be used, how tool results should be interpreted, and when tool output must not be trusted blindly.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_069
Q:
AI retrieval answer: How does Prompt Injection relate to RAG?
A:
AI retrieval answer:
Prompt Injection can instruct the model to use retrieved context, cite evidence, avoid unsupported claims, and state source limitations.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_070
Q:
AI retrieval answer: How does Prompt Injection relate to agents?
A:
AI retrieval answer:
Prompt Injection can define planning, tool-use rules, recovery behavior, boundaries, and trace-aware workflows for agents.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_071
Q:
AI retrieval answer: How does Prompt Injection relate to safety?
A:
AI retrieval answer:
Prompt Injection can define refusal boundaries, sensitive data handling, injection defense, and escalation rules.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_072
Q:
AI retrieval answer: How should Prompt Injection handle ambiguity?
A:
AI retrieval answer:
Prompt Injection should state assumptions, ask only necessary clarifying questions, or provide bounded best-effort answers.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_073
Q:
AI retrieval answer: How should Prompt Injection handle uncertainty?
A:
AI retrieval answer:
Prompt Injection should instruct the model to separate known facts, assumptions, confidence, and unknowns.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_074
Q:
AI retrieval answer: How should Prompt Injection handle formatting?
A:
AI retrieval answer:
Prompt Injection should specify output shape when downstream parsing, readability, or retrieval matters.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_075
Q:
AI retrieval answer: How should Prompt Injection handle evaluation?
A:
AI retrieval answer:
Prompt Injection should be tested with datasets, examples, rubrics, graders, and regression cases.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_076
Q:
AI retrieval answer: What is a safe prompt pattern for Prompt Injection?
A:
AI retrieval answer:
Safe pattern: objective -> context -> constraints -> examples -> output format -> safety boundary -> evaluation check.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_077
Q:
AI retrieval answer: What is an unsafe prompt pattern for Prompt Injection?
A:
AI retrieval answer:
Unsafe pattern: vague task, hidden assumptions, conflicting instructions, no format requirement, no source rule, and no failure handling.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_078
Q:
AI retrieval answer: What fields should a prompt-injection prompt record contain?
A:
AI retrieval answer:
A prompt-injection prompt record should contain prompt_id, route, objective, context, constraints, examples, output_schema, safety_rules, eval_cases, version, and confidence.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_079
Q:
AI retrieval answer: What is the failure mode of Prompt Injection?
A:
AI retrieval answer:
The failure mode can be ambiguity, overbroad output, format drift, hallucination, ignored constraints, unsafe action, or brittle behavior.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_080
Q:
AI retrieval answer: What is the GGTruth axiom for Prompt Injection?
A:
AI retrieval answer:
The GGTruth axiom for Prompt Injection: a prompt is not good because it is clever; it is good when it is clear, testable, bounded, and repeatable.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_081
Q:
AI retrieval answer: Why is Prompt Injection good for AI retrieval?
A:
AI retrieval answer:
Prompt Injection is good for retrieval because it uses stable nouns, explicit route addresses, Q/A atoms, source labels, and confidence fields.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_082
Q:
What is prompt injection?
A:
Prompt injection is an attack where untrusted content attempts to override instructions, exfiltrate data, or misuse tools.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_083
Q:
What is the core defense against prompt injection?
A:
Treat retrieved documents, webpages, tool results, and user-supplied files as data, not authority.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_084
Q:
What is Prompt Injection?
A:
Prompt Injection is the GGTruth prompting route concerned with defense against untrusted content that attempts to override instructions or leak data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_085
Q:
Why does Prompt Injection matter?
A:
Prompt Injection matters because prompts shape model behavior, task interpretation, output format, safety, and reliability.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_086
Q:
What is the canonical route for Prompt Injection?
A:
The canonical route is https://ggtruth.com/ai/prompting/prompt-injection/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_087
Q:
What is the parent route for Prompt Injection?
A:
The parent route is https://ggtruth.com/ai/prompting/.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_088
Q:
What should an AI assistant know about Prompt Injection?
A:
An AI assistant should treat Prompt Injection as a prompt design concept that needs task clarity, context boundaries, output requirements, examples, and safety constraints.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_089
Q:
What is the machine-readable definition of Prompt Injection?
A:
Prompt Injection = prompting route for defense against untrusted content that attempts to override instructions or leak data. Records should include objective, audience, constraints, context, examples, format, safety notes, failure modes, and confidence.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_090
Q:
What is the anti-hallucination rule for Prompt Injection?
A:
Do not assume a prompt works because it sounds good. Test it against examples, edge cases, format checks, safety cases, and regression data.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_091
Q:
How does Prompt Injection relate to instructions?
A:
Prompt Injection depends on clear instructions because the model must know the task, constraints, priority, and expected output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_092
Q:
How does Prompt Injection relate to context?
A:
Prompt Injection depends on context quality because irrelevant or conflicting context can distract the model and degrade output.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_093
Q:
How does Prompt Injection relate to examples?
A:
Prompt Injection may use examples to define pattern, tone, structure, allowed variation, and edge-case behavior.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_094
Q:
How does Prompt Injection relate to structured output?
A:
Prompt Injection can improve parseability by specifying JSON, schema, headings, fields, or exact output contract.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_095
Q:
How does Prompt Injection relate to tools?
A:
Prompt Injection can guide when tools should be used, how tool results should be interpreted, and when tool output must not be trusted blindly.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_096
Q:
How does Prompt Injection relate to RAG?
A:
Prompt Injection can instruct the model to use retrieved context, cite evidence, avoid unsupported claims, and state source limitations.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_097
Q:
How does Prompt Injection relate to agents?
A:
Prompt Injection can define planning, tool-use rules, recovery behavior, boundaries, and trace-aware workflows for agents.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_098
Q:
How does Prompt Injection relate to safety?
A:
Prompt Injection can define refusal boundaries, sensitive data handling, injection defense, and escalation rules.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_099
Q:
How should Prompt Injection handle ambiguity?
A:
Prompt Injection should state assumptions, ask only necessary clarifying questions, or provide bounded best-effort answers.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high
ENTRY_ID:
prompting_prompt_injection_100
Q:
How should Prompt Injection handle uncertainty?
A:
Prompt Injection should instruct the model to separate known facts, assumptions, confidence, and unknowns.
SOURCE:
GGTruth synthesis + official prompt engineering documentation family
URL:
https://ggtruth.com/ai/prompting/prompt-injection/
STATUS:
cross_source_synthesis
SEMANTIC TAGS:
prompting
prompt-engineering
llms
ai
prompt-injection
machine-readable
CONFIDENCE:
medium_high