tk-handoff
Pass
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads from local files (
.tigerkit/seed.mdand.tigerkit/handoff.md) which could contain untrusted data if the repository is compromised. This is a common attack surface for skills that process project documentation. - Ingestion points: Instructions in
SKILL.mdto read task contracts and progress snapshots from the.tigerkit/directory. - Boundary markers: Present; the skill defines a specific text schema (e.g.,
Goal/Seed:,Status:,Decisions:) to structure the data it processes. - Capability inventory: File system read/write access (restricted to the local directory) and shell command execution (limited to Git repository inspection).
- Sanitization: Absent; the skill relies on drift detection and 'verified' status markers rather than explicit content sanitization.
Audit Metadata