ci-integration
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation suggests installing dependencies via 'npx skills add jwilger/agent-skills', which involves downloading and executing code from the author's own repository.\n- [COMMAND_EXECUTION]: The skill performs automated environment interactions, such as running project formatters or linters and executing Git commands (commit and push) to resolve specific build failures.\n- [PROMPT_INJECTION]: The skill ingests full CI logs and build results for classification, introducing a surface for indirect prompt injection (Category 8). 1. Ingestion points: CI log output read during triage (SKILL.md). 2. Boundary markers: No specific markers or delimiters are defined to isolate log data from agent instructions. 3. Capability inventory: Local command execution for linters, Git operations (push/commit), and skill installation via npx. 4. Sanitization: No sanitization or validation of the ingested CI log content is mentioned.
Audit Metadata