salvage
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze untrusted data from Git repository metadata, which could be used to influence agent behavior.
- Ingestion points: The agent is directed to read branch names, commit subjects, diffstats, and external ticket descriptions from issue trackers (SKILL.md).
- Boundary markers: The instructions do not define specific delimiters or "ignore embedded instructions" wrappers for the data read from the repository.
- Capability inventory: The skill grants the agent capabilities to perform destructive actions (branch and worktree deletion) and additive actions (pushing to remotes, creating new branches) based on the classification of this ingested data (SKILL.md).
- Sanitization: There is no mention of sanitizing or escaping the commit messages or branch names before they are processed or summarized in the final report.
Audit Metadata