ot-implement

Pass

Audited by Gen Agent Trust Hub on Nov 15, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from project files such as intent.md, spec.md, and tasks.md. It significantly mitigates injection risks by requiring a manual approval marker (status: approved) in file frontmatter before execution. Additionally, it explicitly instructs the agent to never run commands taken from third-party or fetched content.
  • Ingestion points: on-track/tracks/NNN-slug/intent.md, spec.md, and tasks.md in the target repository.
  • Boundary markers: Mandatory status: approved frontmatter; strict file-level edit restrictions defined in the spec.
  • Capability inventory: File system modification, subprocess execution of validation commands, and execution of local vendor scripts.
  • Sanitization: Relies on human-in-the-loop approval of planning artifacts and a structured post-implementation review.
  • [COMMAND_EXECUTION]: The skill executes validation commands defined within the approved tasks.md and runs local Bash scripts (on-track-gate.sh, on-track-index.sh) provided by the vendor. These scripts are documented as having no network access and are used for repository indexing and quality gating.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 15, 2026, 04:16 AM