handoff
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes common CLI tools (
git,pwd,basename) to extract project metadata, branch information, and uncommitted changes. These commands are executed locally to gather context and do not pose a security risk in this context. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local repository (such as git logs, diffs, and filenames) to synthesize a handoff prompt. This creates an attack surface where malicious content in the project history could influence the generated summary, though the risk is low as the output is intended for user review before being passed to a new session.
- Ingestion points: Reads project state via
git log,git status, andgit diff(SKILL.md). - Boundary markers: Uses Markdown headers as structural delimiters but lacks explicit instructions to the subsequent agent to ignore instructions embedded in the summarized data.
- Capability inventory: The skill is limited to read-only shell commands and does not perform file writes or network operations.
- Sanitization: No sanitization or filtering of project content is performed before interpolation into the output template.
Audit Metadata