skills/othmanadi/plan-it/plan-it-zh/Gen Agent Trust Hub

plan-it-zh

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes lifecycle hooks in the YAML frontmatter (UserPromptSubmit, PreToolUse, PostToolUse, Stop, PreCompact) to run shell commands. These commands are responsible for identifying the local Python interpreter and executing helper scripts like plan-hook.py from the plugin's installation directory to synchronize the planning state.
  • [COMMAND_EXECUTION]: The documentation provides manual commands for the user to execute local scripts, such as scripts/session-catchup.py and scripts/init-plan.sh, to initialize or restore the plan's context.
  • [PROMPT_INJECTION]: To mitigate risks associated with indirect prompt injection from the generated planning artifact (plan.html), the skill implements data boundary markers (===BEGIN PLAN DATA===) and provides an optional SHA-256 attestation feature (/plan-attest) to verify the file's integrity before its content is injected into the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 05:44 AM
Security Audit — agent-trust-hub — plan-it-zh