write-specs
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from repository files, documentation, and code to gather facts for specification documents, creating a potential vector for indirect prompt injection.
- Ingestion points: The 'Fact Gathering' section in SKILL.md describes dispatching sub-agents to look up repository information, documents, and code facts.
- Boundary markers: The skill uses a structured template (references/spec-sdd-template.md) to organize the gathered information, but it lacks explicit isolation delimiters or instructions to ignore embedded commands within the ingested content.
- Capability inventory: The skill has authorization to write SPEC files to the local .specs/ directory as specified in the Overview section.
- Sanitization: No specific sanitization or filtering of the gathered repository content is mentioned or implemented in the instructions.
Audit Metadata