the-loop
01Plugin Claude Code
Plugin marketplace listing 1 plugin: the-loop.
Plugin Claude Code
Plugin marketplace listing 1 plugin: the-loop.
Plugin Claude Code
An opinionated product-development-lifecycle harness that lets coding agents deliver work items end-to-end, with minimal human intervention, through a predictable loop of plan, execute, self-review, critic-review and escalate.
Settings file Claude Code
Agent settings declaring 72 allowed tools.
Instructions file
Instructions for MadaraUchiha-314/the-loop, covering claude.md — working in the-loop's own repository, why this file exists and what to do.
Command
Start a work item with a brainstorm.md scratchpad BEFORE requirements — the optional root artifact. Iterate on it, then convert to requirements with /new-requirement.
Command
Join an EXISTING, in-progress work item (issue or PR) as a contributor — walk the contribution loop toward a human-stated goal and success criteria, without the full spec chain.
Command
Create design.md for a work item from its approved requirements.md (Phase 2 of the spec chain).
Command
Create tasks.md (a DAG of tasks) for a work item from its approved requirements.md, design.md and testing-plan.md.
Command
Create testing-plan.md for a work item from its approved requirements.md and design.md — how the item will be proved, before the task DAG.
Command
Create a work-item ticket (GitHub issue / Jira) from an existing requirements.md, then promote its draft spec folder to docs/specs/ /.
Command
Do an ad-hoc, tactical task on a work item with NO PDLC process — no spec chain, no phase gates, no review chain. Work, ask follow-ups on the thread, continue until the requester says it is done.
Command
Execute the task DAG for a work item — implement, verify against the testing plan, self-check, self/critic-review (implementation → verification → review).
Command
Finalize a work item after all tasks are done — cleanup (currently closing the ticket) and mark the work complete.
Command
Initialize "the-loop" in the current repository — scaffold .the-loop/, the docs trees (specs, capabilities, decisions, learnings) and a validated config, establishing the config with the user via a guided, schema-driven onboarding. Idempotent, non-clobbering, with drift detection.
Command
Start a new work item by drafting a requirements.md BEFORE a ticket exists — in a temporary spec folder. Follow with /create-ticket.
Command
Review a pull request against an authorized reviewer's brief — answer their questions, examine their angles, run their validations, and converse until they say done. Change NO code.
Command
Reconcile a project's "the-loop" files with the installed plugin version — create missing files and migrate schemas. Idempotent, non-clobbering, with --dry-run.
Command
Execute a work item's testing-plan.md after implementation — run the planned activities, record results, commit the evidence (verification phase).
Command
Run "the-loop" on a work item (GitHub issue or Jira id) end-to-end — requirements → design → testing plan → tasks → execute → verify, with self/critic review.
Command
Report the current status of a work item by reading its spec files (requirements/design/tasks) and execution log. Read-only.
Command
Ask a human a question on the work item — the way a spawned agent escalates.
Command
Operate the communication channels — the back-and-forth surfaces beside the work item, starting with the Slack bot (issue-245).
Command
Evaluate a work item's process-graph nodes against its checked-in artifacts, and report what is unmet.
Command
The seam by which the harness doing the work hands it to a different harness for a critic round, and reads back what it said.