agent-introspection

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill emphasizes security best practices, explicitly instructing the agent to ensure no secrets (API keys, tokens) are present in the code or output before submission.
  • [SAFE]: It promotes safe file operations by suggesting the use of a 'trash' directory instead of the destructive 'rm' command.
  • [COMMAND_EXECUTION]: The skill recommends a specific shell command (python3 -c "import json; ...") for the purpose of validating JSON structure. This is a standard utility pattern for ensuring data integrity.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection as it ingests untrusted user feedback and memory data during 'Post-mortem' and 'Weekly Introspection' phases to generate internal 'LEARNINGS'.
  • Ingestion points: Processes user feedback and memory search results in SKILL.md (sections 4 and 5).
  • Boundary markers: Absent; the skill does not explicitly use delimiters when processing these inputs.
  • Capability inventory: Limited to JSON validation via python3 and memory retrieval/recording.
  • Sanitization: Not explicitly defined, though the skill's purpose is to improve logical alignment and accuracy.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 08:01 PM
Security Audit — agent-trust-hub — agent-introspection