skills/uinaf/intake/uinaf-intake/Gen Agent Trust Hub

uinaf-intake

Pass

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a shell script (scripts/publish-entry.sh) to manage a Git workflow, including creating worktrees, running validation via bun, and pushing commits. The operations are scoped to the uinaf/intake repository and follow standard automation practices.
  • [DATA_EXFILTRATION]: While the skill publishes data to a public GitHub repository, the instructions explicitly restrict the intake to public source material only and include specific prohibitions against publishing credentials, private context, or personal data.
  • [EXTERNAL_DOWNLOADS]: The publishing script executes bun install --frozen-lockfile within the repository worktree to ensure validation tools are available. This is standard behavior for Node.js-based project automation.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external public sources (articles, videos, tweets). It includes safety boundaries instructing the agent to summarize in its own words and not reproduce source text directly, which mitigates risks from instructions embedded in those external sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 26, 2026, 11:20 AM
Security Audit — agent-trust-hub — uinaf-intake