understand-change
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data that could contain malicious instructions or hidden payloads.
- Ingestion points: In Step 2 of
SKILL.md, the agent is instructed to read git diffs, touched files, commit messages, and pull request descriptions. - Boundary markers: There are no explicit instructions to the agent to treat this ingested data as untrusted or to ignore any instructions embedded within the git metadata.
- Capability inventory: The agent can execute shell commands via
git(git status,git diff) and perform file writes to the local.turbo/directory. - Sanitization: The skill instructions do not specify any sanitization or filtering of the content read from external git sources before it is used to generate the tutoring checklist.
Audit Metadata