project-planner

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell commands (test and cat) within the Pre-Planning Protocol of SKILL.md to verify the existence of and read content from local project tracking files (docs/plan.md).
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests project data from local files into the agent's context without utilizing boundary markers or sanitization, which could potentially allow data from modified plan files to influence agent behavior.
  • Ingestion points: Content from docs/plan.md and active plan files in docs/plans/ is read into the context using shell commands.
  • Boundary markers: There are no instructions to use delimiters or specific directives to treat the ingested content as untrusted data.
  • Capability inventory: The skill facilitates file system reads/writes and shell command execution.
  • Sanitization: No validation or filtering logic is defined for the content of the project plans before they are processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 03:35 AM
Security Audit — agent-trust-hub — project-planner