spec-architect
Fail
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: CRITICALINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted user input (feature ideas) and use it to generate structured documentation files within the
.agent/specs/directory. This creates an indirect prompt injection surface where a malicious user could potentially embed instructions into a feature request that might influence how the agent interprets or processes the generated documents in subsequent steps. - Ingestion points: Processes user ideas and "vague functional ideas" to fill technical templates in SKILL.md (Phase 1, 2, and 3).
- Boundary markers: Absent. The skill does not provide specific delimiters or instructions to treat user-provided feature descriptions as non-executable data.
- Capability inventory: File-write capabilities to the local
.agent/specs/directory (SKILL.md). - Sanitization: Absent. The skill fills templates directly with user content without escaping or validation logic.
Recommendations
- CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
Audit Metadata