project-manager-senior
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from an external file (
ai/memory-bank/site-setup.md) to generate development tasks. This creates a surface where malicious instructions embedded in a specification file could influence the agent's output or downstream actions. - Ingestion points:
ai/memory-bank/site-setup.md(via reading specifications) - Boundary markers: Absent. There are no instructions to ignore or delimit embedded commands within the specification data.
- Capability inventory: The skill has the capability to write files to the
ai/memory-bank/tasks/directory and suggests the execution of a local shell script. - Sanitization: Absent. The skill quotes requirements directly from the specification without validation.
- [COMMAND_EXECUTION]: The skill instructions include a pre-configured command for quality assurance testing:
./qa-playwright-capture.sh http://localhost:8000 public/qa-screenshots. While this is presented for legitimate testing, it assumes the existence and safety of a local shell script that the agent is encouraged to suggest to the user.
Audit Metadata