resource-development
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external 'Contract' to generate 'Work Packages' and 'Execution Plans'.
- Ingestion points: The 'Contract' document, which acts as the primary input for the decomposition process in Step 1.
- Boundary markers: No specific delimiters or instructions are provided to the agent to distinguish between the skill's instructions and potentially malicious content within the Contract.
- Capability inventory: The skill involves writing to the local file system (e.g.,
docs/gangsta/paths), performing Git operations (creating branches and worktrees), and generating shell commands for 'Verification'. - Sanitization: There is no mention of escaping, filtering, or validating the requirements extracted from the Contract before they are incorporated into task titles, descriptions, or verification commands.
- [COMMAND_EXECUTION]: The workflow requires the agent to generate 'Verification' commands for each work package. If the source Contract contains adversarial instructions designed to influence command generation, it could result in the creation of harmful shell commands that an agent or user might subsequently execute.
Audit Metadata