adr-review
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill focuses on auditing documentation files (ADRs, RFCs, and glossaries) and comparing them against the current implementation. Its operations are restricted to reading project files and providing recommendations to the user.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted data from the repository (e.g., ADR files, design records, and source code). However, it operates as a propose-only tool by default, requiring explicit user approval before making any modifications, which mitigates the risk of the agent being subverted into performing unauthorized actions.
- Ingestion points: Files located in
docs/adr/,docs/fdr/,docs/rfc/,docs/design/,docs/features/, and various source code directories. - Boundary markers: Not explicitly defined in the prompts, though the review is structured as a comparative audit pass.
- Capability inventory: Uses local file reads and read-only shell commands for searching; can propose edits but requires user confirmation to apply them via external skill workflows.
- Sanitization: None observed for processed text content.
Audit Metadata