skill-package-manifest-enricher
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/check.pyusessubprocess.runto execute internal validation tools (validate_skill.py,validate_evals.py) and the manifest enricher. These calls use specific local paths and are restricted to the repository context. - [INDIRECT_PROMPT_INJECTION]: The skill processes frontmatter from
SKILL.mdand content fromevals/evals.json. While these represent external data ingestion points, the script performs standard JSON and string parsing to extract metadata fields like name, description, and test counts without executing the content or using it to construct unsafe prompts. - [DYNAMIC_EXECUTION]: The skill generates a
manifest.enriched.jsonfile at runtime. This process is deterministic based on local metadata and does not incorporate untrusted executable code or remote data. - [DATA_EXPOSURE_AND_EXFILTRATION]: The skill reads repository-level metadata. All data remains within the local filesystem; there are no network calls or patterns suggesting data exfiltration.
Audit Metadata