git-summarize-weekly
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data which could potentially contain malicious instructions aimed at manipulating the AI agent's summary output.
- Ingestion points: Public GitHub events data retrieved via the
gh api users/zc277584121/eventscommand inSKILL.md. - Boundary markers: Absent; there are no clear delimiters or instructions telling the model to ignore formatting or text instructions embedded within the commit messages or PR text.
- Capability inventory: The skill uses the
Bashtool to fetch data, but it only performs text summarization and formatting operations without executing or evaluating the fetched data as code. - Sanitization: Absent; the skill passes the retrieved data directly to the language model for analysis and summarization without pre-filtering or sanitizing text inputs.
Audit Metadata