recall-reasoning
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Python and shell scripts to perform searches.
- Evidence:
uv run python scripts/core/artifact_query.pyandbash "$CLAUDE_PROJECT_DIR/.claude/scripts/search-reasoning.sh". - Context: These scripts appear to be internal tools for managing project state and reasoning history.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from previous sessions which could contain instructions from external sources if they were part of the previous task context.
- Ingestion points: Reads handoffs, plans, and build reasoning logs via
artifact_query.pyandsearch-reasoning.sh. - Boundary markers: None specified in the instructions.
- Capability inventory: Can execute local Python scripts and shell commands described in SKILL.md.
- Sanitization: No explicit sanitization or filtering of recalled text is mentioned.
Audit Metadata