Venture Delegation

Opus-level strategic decomposition for any opportunity, project, or task. Breaks work into atomic pieces with evals, assigns each to the cheapest capable mod...

This skill appears coherent, but keep these practical cautions in mind before enabling it: 1) It expects the agent to create files and run build/eval commands (npm, curl, shell tests). Only allow it where you trust those actions — sandbox or restrict filesystem and network access if possible. 2) Generated atoms may run project scripts that execute code. Review atoms and eval commands before permitting execution. 3) The skill routes work to different internal model names (flash, sonnet, etc.) which can incur costs — verify model availability and cost policies. 4) No secrets are requested by the skill, but if you attach it to a workspace with credentials or CI/CD, ensure it cannot access unrelated secrets. 5) If you want tighter control, require manual approval of the atomic task plan before execution.

Type: OpenClaw Skill Name: venture-delegation Version: 1.0.0 The skill bundle defines a complex delegation framework in SKILL.md that instructs the agent to generate and execute shell commands (e.g., 'npm run build', 'curl', 'grep') to verify task completion. While these capabilities are plausibly needed for the stated purpose of project automation and 'atomic' task verification, the broad use of shell execution and network access represents a high-risk behavior and a significant attack surface for command injection, fitting the criteria for a suspicious classification despite the lack of explicit malicious intent.