gstack-openclaw-retro
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONPERSISTENCECOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests commit messages and author names from the repository history which are contributor-controlled inputs.
- Ingestion points: Commit subjects and author names extracted via git log commands in SKILL.md.
- Boundary markers: No delimiters are used to isolate untrusted commit data from the agent's instructions.
- Capability inventory: The skill executes shell commands and writes files to the local directory.
- Sanitization: No sanitization is performed on commit text before analysis.
- [DATA_EXFILTRATION]: The skill retrieves the local user's name and email through git configuration commands, exposing personal identity data to the session context.
- [PERSISTENCE]: The skill writes JSON files to the 'memory/' folder to track retrospective trends across sessions, creating persistent artifacts on the file system.
- [COMMAND_EXECUTION]: The skill runs extensive git commands to perform log analysis and metric calculations.
Audit Metadata