domains / families / agent-tooling / goal-loop

goal-loop

combined

Draft /goal prompts and explain persistent agent loops. Use for goal or Ralph loops, autonomous run setup, monitoring, and troubleshooting. Draft and explain persistent goal-loop prompts for long-running agent work with clear stop conditions.

When to use

• Use for repeated autonomous work with a **verifiable stop condition**: passing tests, target coverage, eval ≥ X, or a green build. For code work, establish a working build and relevant checks first. No minimum task duration applies. Fits:…
• Task is >30 min of mechanical work.
• There's a **verifiable stop condition** (tests pass, coverage hit, eval ≥ X, build green).
• Repo is agent-ready (working build, decent tests, `AGENTS.md` present).

When not to use

• Adapted from `davidondrej/skills`; verify local paths, tools, credentials, and agent features before acting.
• For commands, remote access, scheduling, browser automation, or file-changing workflows, get explicit user approval and confirm the target environment first.

Skill page

Conflict

The sickn33 copy is a bulk mirror of the davidondrej skill (it explicitly lists source_repo: davidondrej/skills), but it is not a faithful reproduction. It adds material constraints and steps the original lacks (a mandatory Documentation contract item, a 4,000-char limit, a >30-minute task threshold, a repo-readiness requirement, and a specific list of supported agents/auth requirements). It also disagrees with the original on the contract structure (5-part vs 4-part) and the validation cadence (after each change vs during work/final). Because the mirror diverges from the source in ways that would hide real alternatives if one were crowned, the copies should be combined.