tasks-execute
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The bootstrap scripts (
bootstrap.shandbootstrap.ps1) download thearchdevCLI installer from the author's official GitHub repository (ArchAstro/archdev). The download is pinned to a specific commit hash for version consistency. - [REMOTE_CODE_EXECUTION]: The bootstrap scripts execute the downloaded installation script to manage the
archdevCLI. This process includes a mandatory integrity check where the downloaded file's SHA256 hash is verified against a hardcoded value before execution. - [INDIRECT_PROMPT_INJECTION]: The skill processes task descriptions, comments, and dependency lists fetched from a remote server, which could contain instructions intended to influence agent behavior.
- Ingestion points: Task content ingested via
archdev tasks show,tasks list, andtasks commentscommands. - Boundary markers: No explicit delimiters are used to separate task instructions from system prompts.
- Capability inventory: The skill can execute shell commands via the CLI, perform file system modifications (task implementation), and interact with Git repositories.
- Sanitization: No explicit sanitization or filtering of task content is mentioned; however, the skill encourages reading the task fully before claiming it.
Audit Metadata