sota-docs-workflow
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze untrusted external data, such as recent pull requests, code review threads, and commit history, which creates a surface for indirect prompt injection.
- Ingestion points: In
SKILL.mdandrules/01-documentation-architecture.md, the agent is guided to audit "docs tree + README + runbooks", "recent PRs and review threads", and "git history, branches, and tags". - Boundary markers: The instructions do not specify the use of delimiters or specific warnings to ignore embedded instructions when the agent reads this untrusted content.
- Capability inventory: The skill encourages the use of various CLI tools including
git,gh,grep,perl, andawkfor its auditing functions. - Sanitization: There is no mention of sanitizing or validating the content retrieved from PRs or commits before the agent processes it.
- [COMMAND_EXECUTION]: The skill provides several shell command one-liners (using
perl,awk,grep, andgit) intended for the agent to use when auditing the quality and structure of a repository. - Evidence: Multiple audit checklists (e.g., in
rules/02-api-reference-changelogs.md,rules/03-code-review-pr-workflow.md, andrules/04-commits-branches-releases.md) contain specific shell commands for the agent to execute to verify document integrity and repository state.
Audit Metadata