tk-qa-sheet
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python utility located at
scripts/render_static_inventory.pyusingpython3to update the static reading projection of generated HTML artifacts. It also invokestk-browser-verifyfor self-checking the generated sheet. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from repository context such as issue comments, PR reviews, and CI logs.
- Ingestion points: Data enters the agent's context through repository analysis and reading of external file content as described in the
WorkflowandRetrieved Evidence Boundarysections ofSKILL.md. - Boundary markers: The skill explicitly instructs the agent to treat retrieved text as data, not authority, preventing external instructions from altering tool permissions or approved scope.
- Capability inventory: The skill can perform filesystem writes in the
.tigerkit/directory and execute local scripts and browser tools. - Sanitization: The skill implements HTML escaping logic within its Python processing script (
scripts/render_static_inventory.py) and its JavaScript template (assets/qa-sheet-template.html) to prevent Cross-Site Scripting (XSS) in the generated output.
Audit Metadata