aidlc-solutions-review
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE]: The skill reads project design documents, requirements, and manifests from the local file system to identify architectural conflicts. These operations are restricted to project-specific directories defined by environment variables.
- [COMMAND_EXECUTION]: Upon user request to fix an issue, the skill triggers the activation of a related local skill (aidlc-design). This is a legitimate cross-skill workflow for automated design editing.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from design files and manifests. While it lacks boundary markers for these inputs, its capabilities are limited to generating markdown reports and updating YAML status fields.
- Ingestion points: design.md, units.md, and aidlc-manifest.yaml.
- Boundary markers: Not implemented in the provided prompt instructions.
- Capability inventory: Local file read; write to architecture-review.md and aidlc-manifest.yaml; activation of aidlc-design skill.
- Sanitization: None specified for ingested markdown or YAML content.
Audit Metadata