write-specs
Pass
Audited by Gen Agent Trust Hub on May 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional, providing a template and workflow for documentation tasks. No scripts or executable files are included.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions define an attack surface by directing the agent to ingest external data (linked docs, code, issues, PRs, and user notes) to generate specifications.
- Ingestion points: Core Workflow step 1 explicitly instructs reading linked documents, codebases, issues, and PRs.
- Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded prompts within the ingested data.
- Capability inventory: The skill identifies the capability to write the resulting content to a file (
SPEC.md). - Sanitization: No explicit sanitization or validation of the ingested external content is mentioned. While this creates a theoretical surface for indirect prompt injection (e.g., a malicious comment in a PR influencing the spec), the risk is minimal as the output is a static documentation file.
Audit Metadata