Agent Orchestration
AdvisoryAudited by Static analysis on Apr 30, 2026.
Overview
No suspicious patterns detected.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If users spawn multiple agents without tracking them, work could continue longer than intended or duplicate other work.
The skill is explicitly about spawning and tracking sub-agents. The artifact includes tracking controls, so this is purpose-aligned, but it still involves autonomous workers that users should monitor.
Track spawned sub-agents until completion. **No orphans.**
Use explicit labels, expected durations, and completion checks; stop or respawn agents only after reviewing their status and output.
A builder sub-agent could modify files or run commands in the user's workspace as part of completing a requested build.
The builder template instructs an agent to create and test code. This is expected for a build-oriented sub-agent, but it can affect the local environment if the target path, dependencies, and approval expectations are not clear.
Run the code/script to verify it works
Set a clear target directory, dependency policy, and approval requirement before allowing generated code, installs, or scripts to run.
Incorrect or sensitive information recorded in learnings could influence later prompts or expose private project details in local notes.
The skill recommends saving learnings and updating templates for future reuse. This is consistent with the skill's purpose, but persistent notes can carry bad assumptions or sensitive details into later tasks.
Log insights to `notes/resources/prompt-library/LEARNINGS.md`
Review persistent learnings before reuse, avoid storing secrets or confidential data, and remove outdated or untrusted instructions.