retro
Pass
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses repository commit history, local configuration files, and analytics data stored in
~/.gstack/. It also utilizes a telemetry tool to log usage events. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted text from git logs, PR titles, and
TODOS.mdfiles which could contain malicious instructions. - Ingestion points: Git commit messages, PR information, and the
TODOS.mdbacklog file. - Boundary markers: Not present for the ingested text.
- Capability inventory: The skill uses
Bash,Read, andWritetools to perform repository analysis and save session snapshots. - Sanitization: No filtering or sanitization of ingested content is implemented.
- [DYNAMIC_EXECUTION]: The preamble block executes shell commands and local binaries to initialize the session, check for updates, and detect the current repository state.
Audit Metadata