gsd-verify-work
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is configured to execute external workflow logic stored in a local markdown file (@C:/Users/nicol/.codex/get-shit-done/workflows/verify-work.md) using the process directive.
- [SAFE]: Several absolute file paths are hardcoded into the skill (e.g., within execution_context and process sections) that reference a specific user directory ('nicol'). This reveals local system metadata and limits the skill's portability, though it does not constitute a direct security vulnerability.
- [SAFE]: The skill's metadata includes a 'risk: safe' field. This self-assessment is treated as non-authoritative data and does not influence the objective security evaluation.
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface. 1. Ingestion points: user-provided arguments in GSD_ARGS and conversational UAT responses. 2. Boundary markers: No delimiters or warnings are used for interpolated data. 3. Capability inventory: The skill can spawn sub-agents (via spawn_agent) and execute local workflow files (via the process section). 4. Sanitization: External content is not escaped or validated before processing. The risk is assessed as minimal due to the localized testing context.
Audit Metadata