spec-graph
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where the agent treats repository documentation ('specs') as 'authoritative' ground truth, which can lead to the unintended execution of instructions embedded within those documents.\n
- Ingestion points: Content is ingested from Markdown files within the repository using tools like
spec_grep,spec_get, andread(defined in SKILL.md).\n - Boundary markers: The instructions in SKILL.md do not define the use of delimiters or 'ignore' instructions to mitigate the risk of malicious content being processed as commands.\n
- Capability inventory: The agent possesses capabilities to create, modify, and delete files using
spec_create,spec_update,spec_delete,write, andedit(defined in SKILL.md).\n - Sanitization: No validation or sanitization of the specification content is performed before the agent aligns its work with the instructions found therein.
Audit Metadata