implement-full-spec
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill makes extensive use of local shell commands via
gitandgh(GitHub CLI) to manage branches, worktrees, commits, and pull request lifecycle operations. - [EXTERNAL_DOWNLOADS]: Retrieves configuration and task data from well-known services including GitHub, ClickUp, Linear, and Jira. These references are part of the core functionality and target official API endpoints.
- [PROMPT_INJECTION]: The skill is subject to potential indirect prompt injection attacks because it ingests and processes untrusted text from external ticket descriptions and pull request comments to guide code implementation.
- Ingestion points: Data is ingested during Phase A (reading parent tickets from ClickUp, Linear, Jira, or GitHub) and Phase C (fetching pull request review comments).
- Boundary markers: The provided prompt templates in
references/prompt-templates.mddo not employ specific delimiters or instruction-isolation techniques for external content. - Capability inventory: The skill possesses significant capabilities, including the ability to execute shell commands, manage multiple local worktrees, and modify repository content.
- Sanitization: The skill lacks rigorous sanitization of the input text before passing it to agent loops for code modification, although it includes warnings about handling control characters in JSON processing.
Audit Metadata