goal-loopGoal-driven execution supervisor for complex multi-step work: keeps an explicit goal ledger, runs execute→validate→repair→revalidate, blocks premature or lazy completion, maintains PROJECT-CHECKPOINT.md, and resumes interrupted work from the latest verified state. Use when the user asks to complete,
Install via ClawdBot CLI:
clawdbot install bz-ai/goal-loopGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Calls external URL not in known-safe list
https://clawhub.ai/user/bz-aiAudited Sep 12, 2026 · audit v1.0
Generated Oct 26, 2026
A product team uses goal-loop to supervise a multi-file web app build. The agent maintains a requirement ledger, runs typecheck/tests/build/runtime checks, repairs failures through revalidation, and blocks the sprint from closing while any P0 item still lacks evidence.
An engineer asks the agent to fix an intermittent bug where partial fixes are tempting. Goal-loop forces reproduction, root-cause diagnosis, regression tests, and negative-case validation before marking the bug DONE, and checkpoints state so interrupted debugging resumes from the last verified step.
A consultancy needs a Word report, Excel model, and PDF summary delivered together. The agent tracks each deliverable as a ledger row, validates file completeness, table data, page truncation, and citation integrity, and surfaces any scope reduction instead of silently dropping a requirement.
An analyst runs a multi-stage data pipeline and the session gets interrupted. Goal-loop reloads PROJECT-CHECKPOINT.md, skips already-verified stages, and revalidates sampled key values and formulas before continuing, so no work is redone and no result is claimed without evidence.
A research lead coordinates agents producing a deep research brief. The goal ledger enforces source credibility, quoted evidence, and the fact-versus-inference boundary, and the exit gate blocks premature summary until every required section has acceptance evidence.
Distribute the skill under the included proprietary license to agencies and enterprises that need supervised agent execution. Buyers pay per seat or per team for a disciplined, resumable execution layer that plugs into their existing agent setup.
Offer goal-loop as a managed overlay for client projects: humans configure the goal ledger and acceptance criteria, agents execute under the supervisor, and deliveries ship with verifiable validation evidence instead of unchecked completion claims.
Sell vertical packs of prebuilt ledgers, validation checklists, and checkpoint conventions for software, Excel/PPT/PDF, and research workflows. Teams adopt a proven discipline quickly without designing acceptance rules from scratch.
💬 Integration Tip
Install goal-loop in projects where a persistent PROJECT-CHECKPOINT.md can live at the repo root, and pair it with the project's own AGENTS.md/spec so validation checks come from real project rules.
Scored Oct 26, 2026
When the user wants to optimize post-signup onboarding, user activation, first-run experience, or time-to-value. Also use when the user mentions "onboarding...
Guides new OpenClaw users by detecting their environment, recommending skills, and generating a personalized markdown setup report without auto-installing an...
AI-powered career development and education planning intelligence. Tracks job market shifts (BLS/LinkedIn/Burning Glass), emerging roles (7 career domains: A...
Codebase Onboarding
This skill should be used when the user asks to "design interview processes", "create hiring pipelines", "calibrate interview loops", "generate interview que...
Post free job ads to 20+ job boards such as LinkedIn, Indeed, Ziprecruiter etc. to receive applicant resumes via email.