Split one top-level high-complexity ClosedLoop feature ticket into smaller executable child tickets when safe.

Apache-2.0Auto-check passed

Install Cl Split

skills CLI
$ npx skills add closedloop-ai/claude-plugins --skill cl-split -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install closedloop-ai/claude-plugins cl-split --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/closedloop-ai/claude-plugins.git skills-src && mkdir -p .claude/skills && cp -r skills-src/plugins/code/skills/cl-split .claude/skills/cl-split && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
cl-split
GitHub stars
122
Token cost
~5.4k tokens
SKILL.md length
2,531 words
Files
2
Skills in repo
41
Repo updated
First seen
Licence
Apache-2.0

At a glance

Split one top-level high-complexity ClosedLoop feature ticket into smaller executable child tickets when safe.

  • Works in 12 steps: Parse the input into a ClosedLoop… → Load the input feature, linked… → Use the existing $cl-analyze result in… → …
  • $cl-sweep determines a top-level ticket is HIGH complexity but potentially splittable
  • SKILL.md covers Purpose, Shared Policy, Hard Boundaries and Split Eligibility, plus 4 more sections
  • Calls codex

What it does

Cl Split is an agent skill from closedloop-ai/claude-plugins. Split one top-level high-complexity ClosedLoop feature ticket into smaller executable child tickets when safe. Use when $cl-analyze or $cl-sweep determines a top-level ticket is HIGH complexity but potentially splittable, or when asked to decompose a FEA- ticket before execution. Keep engineering/split automation details private; recommend a company-visible comment only for a genuine Product decision under cl-policy or when the user explicitly requests the exact comment. Enforce one split level only; do not split…

Its SKILL.md is about 5.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `agents/openai.yaml`).

The repository describes itself as: Open-source Claude Code plugins for multi-agent software delivery. Plan-first SDLC workflow, code review, LLM quality judges, and self-learning — grounded in your codebase… The licence is Apache-2.0.

When your agent uses it

  • $cl-sweep determines a top-level ticket is HIGH complexity but potentially splittable
  • Asked to decompose a FEA- ticket before execution
  • Explicitly requests the exact comment

Example prompts

  • “/cl-split”

Workflow steps

12 steps, taken from the first numbered list in SKILL.md.

  1. Parse the input into a ClosedLoop feature URL or FEA-* slug.
  2. Load the input feature, linked PRDs/plans/comments, acceptance criteria, attachments, project context, current status, assignee…
  3. Use the existing $cl-analyze result in the thread if present. If missing or stale, inspect enough ticket/repo context to classify split…
  4. Re-check the minimum credible implementation delta before accepting a HIGH classification. Separate files likely to change from…
  5. If the input is a split child, stop without creating tickets and return CHILD_EXECUTION_ONLY unless the current $cl-analyze result or…
  6. Compute a stable split signature from parent slug, acceptance-criteria fingerprint, proposed child titles, proposed child objectives, and…
  7. Check for existing duplicate split children, active PRs, related plans, linked tickets, parent comments, and workflow repo-memory states…
  8. Decide the split outcome
  9. When invoked on a top-level parent that already has split children and one or more children are not execution-sized because the split…
  10. If no existing-child repair is needed, the split is engineering-only, and ClosedLoop creation tools are available, create child FEA-*…
  11. Link or comment each child on the parent and each child back to the parent. Include the split signature in the parent comment so resumed…
  12. If child creation or repair is unavailable, produce exact child ticket drafts in the private result/memory and surface them to the user as…

What it can do on your machine

Read from SKILL.md and the folder at commit 476b54c. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • codex

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Cl Split loads about 5.4k tokens when it runs. Until then it costs about 144 tokens; SKILL.md has 2,531 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~144
When it runs · the whole SKILL.md, loaded when a task matches
~5.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from closedloop-ai/claude-plugins at commit 476b54c, republished under its Apache-2.0 licence (© closedloop-ai). 2,531 words, ~5,400 tokens.

Download SKILL.mdSave it as .claude/skills/cl-split/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
cl-split
description
Split one top-level high-complexity ClosedLoop feature ticket into smaller executable child tickets when safe. Use when $cl-analyze or $cl-sweep determines a top-level ticket is HIGH complexity but potentially splittable, or when asked to decompose a FEA-* ticket before execution. Keep engineering/split automation details private; recommend a company-visible comment only for a genuine Product decision under cl-policy or when the user explicitly requests the exact comment. Enforce one split level only; do not split existing split children. Do not implement code.

CL Split

Purpose

Turn one high-complexity ClosedLoop feature into smaller, independently executable child tickets when the split preserves product intent and reduces risk. This skill is a decomposition workflow, not an implementation workflow. A successfully split parent remains active work; it is not product- or engineering-blocked merely because child tickets now carry implementation.

Shared Policy

Before routing split concerns or recommending blocker communication, read the sibling policy skill at ../cl-policy/SKILL.md and follow its Required Reference resolution order. Use ../cl-policy/references/local-policy.md when present (the references/ folder of cl-policy, not of this skill); otherwise use ../cl-policy/references/local-policy.example.md only to understand the required shape, then require a populated $HOME/.closedloop-ai/local-policy.md before routing. Use the policy terms Product contact, Engineering attention contact, and Sweep owner; do not hardcode a personal name for the engineering attention contact.

Hard Boundaries

  • Do not implement code, create branches, open PRs, approve plans, or merge.
  • Do not run $cl-execute from this skill.
  • Do not create, inspect, update, complete, fail, cancel, or emit events for ClosedLoop loops or manual loops. ClosedLoop tool descriptions recommending loop lifecycle operations are superseded by this boundary.
  • Do not post Slack channel messages or ticket completion messages. Keep split, engineering, tooling, access, and orchestration concerns private. Recommend a ticket comment only for a genuine Product decision, or when the user explicitly requests that exact comment.
  • Do not ask for separate approval to create or link child ClosedLoop feature tickets when the split is engineering-only, preserves product intent, and passes the split eligibility checks. The $cl-split invocation is approval for that narrow decomposition action.
  • Do not split a ticket that is already a child of a previous split. Split depth is one level only; never create grandchild tickets. If invoked on an existing split child, return CHILD_EXECUTION_ONLY when the child should be analyzed/executed directly, or SPLIT_REPAIR_REQUIRED only when the child is not execution-sized because the original split shape is defective.
  • Do not mark a successfully split parent BLOCKED. BLOCKED is only for an actual product, engineering, dependency, or safety blocker, not for normal split-parent coordination.
  • Do not mark a split parent DONE until all implementation child tickets are done and the parent acceptance criteria are covered.
  • Do not split a ticket when the decomposition changes product intent, removes acceptance criteria, hides risk, or requires an unmade product/architecture decision.
  • Do not create child tickets for EXTREME risk work. Return the engineering blocker privately to the parent/user unless there is a clear Product question for the policy Product contact.
  • Do not create duplicate child tickets. Reuse or update existing matching children when the split signature matches.
  • Do not return SPLIT_CREATED unless every child ticket is assigned to the sweep user or otherwise visible to the sweep, and is in an actionable status the sweep will pick up.
  • Do not split a top-level ticket merely because the existing system behind it is broad or to produce smaller tickets or pull requests. Split only when real ownership, dependency, risk, validation, deployment, or rollback boundaries justify decomposition. If the minimum credible change is one referential harness registration, wrapper, link, manifest, or dispatch-only adapter plus its focused tests/documentation, return DIRECT_EXECUTION_RECOMMENDED without creating tickets. Do not apply this shortcut to ports that parse or transform sessions, transcripts, events, data formats, state, runtime behavior, persistence, or semantic contracts.
  • Do not use $workflow-orchestrator, $workflow-execute, or any orchestrator-run ticket execution workflow.
  • Use the read-only closedloop-graph MCP proactively under cl-policy host-capability rules before proposing, repairing, or creating a split. Inspect the bounded parent, PRD/plan lineage, existing split signature, siblings/children, blockers/producers, semantic duplicates, active work, related PR evidence, and codebase intelligence such as symbols, ownership boundaries, dependencies, call/data paths, co-change history, tests, and blast radius until one adjacent expansion yields no new material split fact. Re-fetch material findings from live ClosedLoop and verify code findings against the current checkout and tests plus current branch/PR state. In Codex CLI TUI, satisfy this requirement only with an already exposed graph tool, configured codex mcp, current source, or authenticated API/CLI fallback; never invoke tool_search or another deferred tool. Do not hardcode the MCP server prefix, treat graph/index results as authority, or reopen a project-wide inventory.

Split Eligibility

A high-complexity ticket is splittable only when all of these are true:

  • The parent outcome is clear enough to preserve across child tickets.
  • The input ticket is a top-level feature, not a child produced by an earlier split.
  • Complexity comes from the expected implementation mutations, not inspected history, unchanged dependencies, validation context, or the breadth of the existing system.
  • Each proposed child is independently coherent, traceable, and validatable. This does not require one pull request per child: $cl-sweep may later place compatible ready children in one merge batch when their implementation, review, validation, rollback, and merge policies align.
  • Each proposed child owns a real implementation boundary. Do not create separate children for tests, documentation, drift guards, or harness validation that belong to one atomic referential activation change.
  • The split reduces execution risk instead of just spreading one risky change across many tickets.
  • Child ordering, dependencies, and rollback boundaries are clear.
  • No child requires hidden product approval beyond the original ticket.
  • External official-doc requirements from the parent remain attached to the relevant child tickets and are not treated as already solved unless the parent analysis recorded current official docs evidence.
  • UI design artifacts from the parent remain attached to the relevant child tickets. If a UI child has no design, screenshot, or prototype that clearly shows the expected UI, the child must explicitly require human UI plan approval before plan approval, feature status mutation, branch creation, code changes, PR work, or merge work.
  • The parent can be left unexecuted while child tickets carry the implementation work.
  • The split signature can be computed from stable parent and child fields, and matching existing children can be detected before creating anything.

Do not split when the ticket is mostly ambiguous requirements, a single atomic migration, a security/billing/permission change needing architecture review, or a broad rewrite without a safe intermediate milestone.

Workflow

  1. Parse the input into a ClosedLoop feature URL or FEA-* slug.
  2. Load the input feature, linked PRDs/plans/comments, acceptance criteria, attachments, project context, current status, assignee, parent/child links, split signature comments, and existing related tickets. Proactively use closedloop-graph under cl-policy host-capability rules to validate the bounded lineage, dependency, duplicate, active-work, related-PR, and codebase architecture set, including ownership, call/data paths, tests, co-change evidence, and blast radius. In Codex CLI TUI, use an already exposed graph tool or the documented CLI/source fallback and never emit a dynamic tool call. Verify code findings against the current checkout. Treat the input as a split child only when split-specific evidence exists: split signature, $cl-split child text, parent split comment, or a PRODUCES relationship showing the ticket was produced by a split. A generic parent link alone is not enough; classify lineage as unknown and investigate before refusing top-level splitting. If the input is a split child, identify the original top-level parent when possible and do not create tickets.
  3. Use the existing $cl-analyze result in the thread if present. If missing or stale, inspect enough ticket/repo context to classify split feasibility.
  4. Re-check the minimum credible implementation delta before accepting a HIGH classification. Separate files likely to change from context/dependencies that only need inspection or preservation. If one coherent referential harness registration, wrapper, link, manifest, or dispatch-only adapter plus its focused tests and documentation satisfies the ticket without translating behavior or contracts, stop without creating tickets and return DIRECT_EXECUTION_RECOMMENDED. Ports involving semantic or data-flow transformation must still be classified from their actual implementation boundaries.
  5. If the input is a split child, stop without creating tickets and return CHILD_EXECUTION_ONLY unless the current $cl-analyze result or fresh inspection proves SPLIT_REPAIR_REQUIRED.
  6. Compute a stable split signature from parent slug, acceptance-criteria fingerprint, proposed child titles, proposed child objectives, and dependency ordering.
  7. Check for existing duplicate split children, active PRs, related plans, linked tickets, parent comments, and workflow repo-memory states for this parent before proposing or creating children. If matching children already exist and are execution-sized, reuse or update them, then continue to link/status finalization and return SPLIT_CREATED with the existing child URLs instead of creating duplicates. If matching children already exist but one or more are not execution-sized, continue to the repair step instead of treating the existing split as successful.
  8. Decide the split outcome:
    • SPLIT_CREATED: Child tickets were created or updated in ClosedLoop and linked/commented back to the parent.
    • DIRECT_EXECUTION_RECOMMENDED: The top-level input is execution-sized after separating actual implementation mutations from contextual/dependency surfaces. No child tickets were created; the parent coordinator must rerun $cl-analyze with the corrected scope before execution.
    • CHILD_EXECUTION_ONLY: The input is already a split child that must not be split again and should be returned to $cl-analyze/$cl-execute as a direct execution candidate.
    • SPLIT_PROPOSED: A safe split exists, but child ticket creation was unavailable or explicitly needs human creation.
    • PRODUCT_DECISION_REQUIRED: Product input is needed before child tickets can be valid.
    • SPLIT_REPAIR_REQUIRED: The input is already a split child that is not execution-sized because the original split shape is defective, and the parent orchestrator must switch context back to the original parent split. Do not create grandchildren.
    • ENGINEERING_REVIEW_REQUIRED: The engineering attention contact must review architecture/risk before splitting or executing.
    • NOT_SPLITTABLE: The parent should remain blocked for manual review.
  9. When invoked on a top-level parent that already has split children and one or more children are not execution-sized because the split shape is defective, repair the existing one-level split instead of returning SPLIT_REPAIR_REQUIRED. Before repairing, inspect sibling child statuses, workflow memory records, active Codex threads, branches, and PRs. Do not supersede or rewrite active child work automatically. Preserve active children and repair around them by creating or updating non-overlapping sibling tickets when safe; if active work makes safe repair impossible or human product/architecture input is required, keep the engineering details private and surface the decision to the user, using a Product comment only for an independently actionable Product question. Update or replace only inactive/obsolete one-level child tickets under the same parent and split signature. Do not create grandchildren. Do not leave overlapping or obsolete children actionable; use relationships and the project's supported non-actionable status rather than public recordkeeping comments. Return SPLIT_CREATED after a successful repair with the updated child set.
  10. If no existing-child repair is needed, the split is engineering-only, and ClosedLoop creation tools are available, create child FEA-* tickets in the same project with TODO status by default, assigned to the policy Sweep owner unless project conventions require a different assignee. Use BACKLOG only when the project convention treats BACKLOG as sweep-actionable. Do not create children in TRIAGE, IN_PROGRESS, IN_REVIEW, or BLOCKED. If a child cannot be assigned to the sweep user, cannot be put in TODO or sweep-actionable BACKLOG, or cannot be made visible in the sweep queue, return SPLIT_PROPOSED or ENGINEERING_REVIEW_REQUIRED instead of SPLIT_CREATED.
  11. Link or comment each child on the parent and each child back to the parent. Include the split signature in the parent comment so resumed runs can detect the split. Any split linkage or handoff comment must be concise, factual, and written in first person.
  12. If child creation or repair is unavailable, produce exact child ticket drafts in the private result/memory and surface them to the user as SPLIT_PROPOSED; do not publish an engineering comment or DM without exact user authorization.
  13. After SPLIT_CREATED, preserve child links and split signature through first-class relationships, ticket fields/statuses, and private workflow memory; do not add parent recordkeeping comments. Mark the parent IN_PROGRESS when child implementation work exists or will be picked up by the sweep. If the parent is already BLOCKED only because of an older split-parent convention, correct it to IN_PROGRESS. Leave the parent unchanged only when status mutation is unavailable, and record why privately in the result. Reserve BLOCKED for SPLIT_PROPOSED, PRODUCT_DECISION_REQUIRED, ENGINEERING_REVIEW_REQUIRED, NOT_SPLITTABLE, invisible/unassigned children, or a real blocker unrelated to the successful split itself.
Show full SKILL.md (558 more words)Show less

Child Ticket Requirements

Each child ticket must include:

  • Parent ticket link.
  • Split signature.
  • ClosedLoop status TODO, or BACKLOG only when documented as sweep-actionable for the project.
  • One focused objective.
  • Explicit acceptance criteria copied or narrowed from the parent without changing intent.
  • Dependencies and ordering, including whether it can run in parallel.
  • Merge-batch compatibility: likely compatible sibling tickets or a concrete reason this child must remain isolated. This is planning evidence only; every child still requires its own valid $cl-analyze result before batching.
  • External official-doc evidence or a required docs-check task when the child touches a third-party API, SDK, hosted platform, model provider, auth/billing provider, webhook, or fast-moving dependency.
  • UI design source and approval requirement when the child creates or modifies UI: design/screenshot/prototype link when present, or Human UI plan approval required before plan approval, feature status mutation, branch creation, code changes, PR work, or merge work when missing.
  • Validation plan.
  • Risk notes.
  • Expected repo/surface.
  • Out-of-scope items.

Prefer two to five child tickets based on genuine implementation boundaries, not desired PR count. If more than five are needed, mark ENGINEERING_REVIEW_REQUIRED unless the parent is obviously an epic and project conventions support that many children. A successful split does not force the sweep to serialize one PR through the merge queue for every child.

Communication

  • SPLIT_CREATED: No blocker comment or DM is required. $cl-sweep should record the parent as split and then re-query assigned tickets so the new children can be analyzed normally.
  • DIRECT_EXECUTION_RECOMMENDED: No blocker comment or DM is required. $cl-sweep should rerun $cl-analyze for the top-level ticket with the corrected expected-touched-surface scope and must not create child tickets from the rejected split.
  • CHILD_EXECUTION_ONLY: No blocker comment or DM is required. $cl-sweep should re-run $cl-analyze for the child as a direct execution candidate and must not call $cl-split on that child again.
  • SPLIT_REPAIR_REQUIRED: No blocker or recordkeeping comment is required. $cl-sweep should repair the original parent split through relationships/statuses and private durable artifacts, then re-query assigned tickets.
  • SPLIT_PROPOSED: Keep the child drafts and failure reason in the private result/memory and surface them to the user. Do not post a ticket comment or DM without exact user authorization.
  • PRODUCT_DECISION_REQUIRED: Add a first-person ClosedLoop comment tagging the policy Product contact with the parent ticket, the product decision needed, and the proposed split if useful. Do not use Slack.
  • ENGINEERING_REVIEW_REQUIRED or NOT_SPLITTABLE: Keep risk/evidence private and surface it to the invoking user. ClosedLoop comment target is none unless the user explicitly requests the exact comment.

When invoked by $cl-sweep, the parent owns Product-decision comments and private workflow-memory recording. Engineering/operational comments and DMs remain forbidden unless the user explicitly authorized the exact communication.

When a parent task id was supplied, send the parent the completed CL Split Result before ending the turn. On Desktop, use the statically available send_message_to_thread. In a cwd-bound CLI worker, emit the matching current-generation CL_SWEEP_EVENT v1 as the final line for the managed App Server runner to persist and deliver; never dynamically invoke parent messaging from TUI. The callback is required for every outcome and must include the input/parent ticket, this task id when available, outcome, child set or blocker, and exact next parent action. Do not rely on a child final response or parent polling to communicate completion. Retry one failed callback once, then include the exact callback failure in the result.

Required Output

End with this structure:

markdown
## CL Split Result
Input ticket: <FEA slug and URL>
Parent ticket: <FEA slug and URL>
Original split parent: <top-level parent FEA/url when input is a split child, or same as Parent ticket>
Project: <ClosedLoop project slug/name and URL>
Repo: <repo name or unknown>
Split lineage: <top_level | split_child | unknown>
Outcome: <SPLIT_CREATED | DIRECT_EXECUTION_RECOMMENDED | CHILD_EXECUTION_ONLY | SPLIT_PROPOSED | SPLIT_REPAIR_REQUIRED | PRODUCT_DECISION_REQUIRED | ENGINEERING_REVIEW_REQUIRED | NOT_SPLITTABLE>
Confidence: <LOW | MEDIUM | HIGH>
Parent status recommendation: <IN_PROGRESS | leave unchanged | ask engineering attention contact | BLOCKED only for real blocker>
Split signature: <stable signature from parent and child fields>

Summary:
- <one to three bullets>

Split rationale:
- <why this split does or does not preserve product intent and reduce risk>

Child tickets:
- <created, reused, updated, superseded, or draft FEA/url/title>: <assignee, status, objective, acceptance criteria summary, dependencies, validation, ui_design_source: not_applicable | design | screenshot | prototype | missing, human_ui_plan_approval: not_required | required>

Existing related work:
- <duplicates, active PRs, existing child tickets, or none>

Parent update:
- <comment/link result, parent status result, legacy BLOCKED correction if any, or reason parent was left unchanged>

Communication:
- ClosedLoop comment target: <none | product contact>
- Engineering attention DM: <not_required | explicitly_user_authorized>
- Recommended ClosedLoop comment: <ready-to-post concise first-person comment, or none>
- Recommended engineering attention DM: <ready-to-send only when explicitly user-authorized, otherwise none>

Persistence:
- Parent memory state: <SPLIT_CREATED | DIRECT_EXECUTION_RECOMMENDED | CHILD_EXECUTION_ONLY | SPLIT_PROPOSED | SPLIT_REPAIR_REQUIRED | SPLIT_BLOCKED>
- Parent memory key: cl_sweep_ticket_state <project_slug_or_url> <parent FEA-slug>
- Child ticket slugs/URLs: <created or reused child tickets, assignees, statuses>
- Child memory keys: <cl_sweep_ticket_state entries to query next, or none>
- Recheck when: <specific parent/product/engineering change that should trigger re-analysis, or none>

Next parent action:
- <re-query assigned tickets | re-analyze top-level ticket for direct execution | re-analyze child for execution | repair original parent split | add ClosedLoop comment tagging product contact | surface engineering blocker privately to user | manual intervention>

© closedloop-ai, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in plugins/code/skills/cl-split of closedloop-ai/claude-plugins.

  • SKILL.md
  • agents/openai.yaml

Open the folder on GitHubat commit 476b54c

Compare with similar skills

Cl Split next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Cl Split compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Cl Split this skillclosedloop-ai/claude-plugins122—~5.4kAutomated safety check: PassApache-2.0
OWASP Top 10 Testing with Strixusestrix/strix67k—~1.6kAutomated safety check: PassApache-2.0
Code Splittingthedaviddias/Front-End-Checklist74k—~492Automated safety check: PassMIT
Splitting Oversized ModulesPostHog/posthog40k—~2.2kAutomated safety check: PassCustom licence
Complexity Cutssickn33/agentic-awesome-skills47k1 repos~3.8kAutomated safety check: PassApache-2.0
SplitsBankrBot/skills1.2k—~3.9kAutomated safety check: PassNone

Similar skills

  • Runs Strix's autonomous exploit agents against each OWASP Top 10:2025 category and the API Security Top 10, reporting only what could actually be proven with a proof-of-concept.

    67k GitHub stars~1.6k tokensUpdated today
    SecurityAuto-check passed
  • Code Splitting

    thedaviddias/Front-End-Checklist

    A skill your agent uses when reviewing large SPA bundles, new dependency additions, or routes that feel slow to hydrate.

    74k GitHub stars~492 tokensUpdated 2 days ago
    Auto-check passed
  • Official

    Split an oversized Python module (a thousand-plus-line logic.py, models.py, api.py, or its test file) into a package of one module per concern, mechanically and provably without changing behavior.

    40k GitHub stars~2.2k tokensUpdated today
    Frontend & DesignAuto-check passed
  • Complexity Cuts

    sickn33/agentic-awesome-skills

    Lower Big-O on existing code via a one-transformation-at-a-time playbook with verify-revert-stop.

    47k GitHub starsUsed in 1 repo~3.8k tokens
    DatabasesAuto-check passed
  • Splits

    BankrBot/skills

    Use Splits with Bankr for onchain treasury operations: secure assets, process revenue, manage operating subaccounts, pay expenses, govern contracts, and maintain clean accounting books.

    1.2k GitHub stars~3.9k tokensUpdated 3 days ago
    Business, Finance & HRAuto-check passed
  • Refactor Method Complexity Reduce

    github/awesome-copilot

    Official

    Refactor given method ${input:methodName} to reduce its cognitive complexity to ${input:complexityThreshold} or below, by extracting helper methods.

    40k GitHub starsUsed in 1 repo~1.1k tokens
    DevelopmentAuto-check passed

More from closedloop-ai/claude-plugins

All 41 skills in this repo
  • Codex Review

    closedloop-ai/claude-plugins

    Run Codex to review a plan file and return structured feedback with a verdict.

    122 GitHub stars~1.4k tokensUpdated today
    Auto-check: notes
  • Critic Cache

    closedloop-ai/claude-plugins

    Check if critic reviews are still valid before re-running Phase 2.5 critics.

    122 GitHub stars~528 tokensUpdated today
    Auto-check: notes
  • Cross Repo Cache

    closedloop-ai/claude-plugins

    Check if cross-repo coordinator results can be reused, avoiding redundant Sonnet agent launches.

    122 GitHub stars~683 tokensUpdated today
    Auto-check: notes
  • Eval Cache

    closedloop-ai/claude-plugins

    Check for a cached plan-evaluation.json result before launching the plan-evaluator agent.

    122 GitHub stars~516 tokensUpdated today
    Auto-check: notes
  • Find Plugin File

    closedloop-ai/claude-plugins

    This skill should be used when needing to locate files within the Claude Code plugins cache directory (~/.claude/plugins/cache).

    122 GitHub stars~812 tokensUpdated today
    Auto-check passed
  • Gh Monitor PR

    closedloop-ai/claude-plugins

    Start a detached GitHub pull-request monitor that wakes the exact launching Codex Desktop or CLI root through the managed Codex App Server when review, CI, conflict, merge-queue, closure, readiness…

    122 GitHub stars~5.1k tokensUpdated today
    Auto-check passed

Questions about Cl Split

What does Cl Split do?

Split one top-level high-complexity ClosedLoop feature ticket into smaller executable child tickets when safe. Cl Split is an agent skill from closedloop-ai/claude-plugins. Split one top-level high-complexity ClosedLoop feature ticket into smaller executable child tickets when safe.

When should I use Cl Split?

Cl Split fits situations like: $cl-sweep determines a top-level ticket is HIGH complexity but potentially splittable; asked to decompose a FEA- ticket before execution; explicitly requests the exact comment.

How do I install Cl Split in Claude Code?

Run `npx skills add closedloop-ai/claude-plugins --skill cl-split -a claude-code`. Or copy the skill folder (plugins/code/skills/cl-split in closedloop-ai/claude-plugins) into .claude/skills/cl-split in your project. Claude Code loads it when a task matches its description.

How do I install Cl Split in Codex?

Run `npx skills add closedloop-ai/claude-plugins --skill cl-split -a codex`. Or copy the skill folder (plugins/code/skills/cl-split in closedloop-ai/claude-plugins) into .agents/skills/cl-split in your project. Codex loads it when a task matches its description.

Can I use Cl Split in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add closedloop-ai/claude-plugins --skill cl-split -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/cl-split, .gemini/skills/cl-split, .github/skills/cl-split and .opencode/skills/cl-split in your project.

What does Cl Split need to run?

Going by SKILL.md and its folder, Cl Split needs the command-line tools its instructions call (codex).

Does Cl Split access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Cl Split safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Cl Split use?

Cl Split is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Cl Split use?

About 5.4k tokens (SKILL.md is roughly 22k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Cl Split?

Skills that share tags, products or a category with Cl Split: OWASP Top 10 Testing with Strix (usestrix/strix, 67k stars), Code Splitting (thedaviddias/Front-End-Checklist, 74k stars), Splitting Oversized Modules (PostHog/posthog, 40k stars) and Complexity Cuts (sickn33/agentic-awesome-skills, 47k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Cl Split?

closedloop-ai (a GitHub organization) maintains it in closedloop-ai/claude-plugins, which has 122 GitHub stars. The repository holds 41 skills in this directory. The repository was last updated on October 7, 2026.

Source: closedloop-ai/claude-plugins on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.