skill-map
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
ai-skilletCLI to map and analyze skills on the local filesystem. - [INDIRECT_PROMPT_INJECTION]: The skill scans local files for skill definitions and references, which could serve as a vector for indirect prompt injection if malicious content is processed.
- Ingestion points: Local files scanned via root or portfolio arguments (SKILL.md).
- Boundary markers: Implements a comprehensive ignore policy (references/ignore-policy.md) for folders like
.git,node_modules, and agent state directories. - Capability inventory: Execution of
ai-skillet mapshell commands (SKILL.md). - Sanitization: Prescribes the use of structured JSON/DOT output formats to maintain data integrity (SKILL.md).
Audit Metadata