skill-refine
Skill4.4k starsChanged 3 months agoMITView on GitHub
Score and improve an artifact with a frozen rubric, independent models, implemented edits, and bounded rescoring
- Deletes or force-pushes
3 other projects use a version of this.
Adopted by
Steady: 0 new projects took it in the last 4 weeks, none the 4 weeks before.thin data
Most recently taken up by
- DentalEdge-Solutions/claude_codetheir own version, first seen Nov 1, 2026
- windgeek/skillstheir own version, first seen Oct 15, 2026
- maxenko/claude-skillstheir own version, first seen Jul 14, 2026
Who's talking about nyldn/claude-octopus
- hashgraph-online/awesome-codex-plugins“Add Claude Octopus — multi-LLM orchestration plugin”GitHub · Mar 29, 2026 · 3 reactions
- @andybrnr.bsky.social“Since individual models have their blind spots, ensembling different coding agents can lead to more robust …”Bluesky · Mar 31, 2026 · 2 likes
- A Hacker News story“Multi-tentacled orchestrator for Claude Code”Hacker News · Mar 18, 2026 · 1 points
- sudosubin/agents.nix“codex-plugins.github.nyldn.claude-octopus: 11.23.0 → 11.24.0”GitHub · Nov 15, 2026
- sudosubin/agents.nix“claude-code-marketplaces.github.nyldn.claude-octopus: 11.23.0 → 11.24.0”GitHub · Nov 15, 2026
The biggest 5 of 53 mentions linking to the repository that our crawl found. Each line is their own headline.
See all 53
Talked about on GitHub, Bluesky and Hacker News (53 times).
On GitHub · 50
- Mar 29, 2026Add Claude Octopus — multi-LLM orchestration plugin — hashgraph-online/awesome-codex-plugins3 reactions · 6 comments
- Nov 15, 2026codex-plugins.github.nyldn.claude-octopus: 11.23.0 → 11.24.0 — sudosubin/agents.nix0 reactions
- Nov 15, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.23.0 → 11.24.0 — sudosubin/agents.nix0 reactions
- Nov 15, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.23.0 → 11.24.0 — sudosubin/agents.nix0 reactions
- Nov 15, 2026agent-skills.github.nyldn.claude-octopus: 11.23.0 → 11.24.0 — sudosubin/agents.nix0 reactions
45 more on GitHub
- Nov 13, 2026agent-skills.github.nyldn.claude-octopus: 11.22.2 → 11.23.0 — sudosubin/agents.nix0 reactions
- Nov 12, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.22.2 → 11.23.0 — sudosubin/agents.nix0 reactions
- Nov 12, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.22.2 → 11.23.0 — sudosubin/agents.nix0 reactions
- Nov 12, 2026codex-plugins.github.nyldn.claude-octopus: 11.22.2 → 11.23.0 — sudosubin/agents.nix0 reactions
- Nov 12, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.22.1 → 11.22.2 — sudosubin/agents.nix0 reactions
- Nov 12, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.22.1 → 11.22.2 — sudosubin/agents.nix0 reactions
- Nov 12, 2026codex-plugins.github.nyldn.claude-octopus: 11.22.1 → 11.22.2 — sudosubin/agents.nix0 reactions
- Nov 12, 2026agent-skills.github.nyldn.claude-octopus: 11.22.1 → 11.22.2 — sudosubin/agents.nix0 reactions
- Nov 11, 2026codex-plugins.github.nyldn.claude-octopus: 11.22.0 → 11.22.1 — sudosubin/agents.nix0 reactions
- Nov 11, 2026agent-skills.github.nyldn.claude-octopus: 11.21.0 → 11.22.1 — sudosubin/agents.nix0 reactions
- Nov 11, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.21.0 → 11.22.1 — sudosubin/agents.nix0 reactions
- Nov 11, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.21.0 → 11.22.1 — sudosubin/agents.nix0 reactions
- Nov 11, 2026codex-plugins.github.nyldn.claude-octopus: 11.21.0 → 11.22.0 — sudosubin/agents.nix0 reactions
- Nov 10, 2026Liga cinco plugins do Claude Code nas sessões do repositório — Leirbagrazilian/Appdeestudos0 reactions · 1 comment
- Nov 10, 2026agent-skills.github.nyldn.claude-octopus: 11.20.0 → 11.21.0 — sudosubin/agents.nix0 reactions
- Nov 10, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.20.0 → 11.21.0 — sudosubin/agents.nix0 reactions
- Nov 9, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.19.0 → 11.21.0 — sudosubin/agents.nix0 reactions
- Nov 9, 2026codex-plugins.github.nyldn.claude-octopus: 11.19.0 → 11.21.0 — sudosubin/agents.nix0 reactions
- Nov 9, 2026agent-skills.github.nyldn.claude-octopus: 11.19.0 → 11.20.0 — sudosubin/agents.nix0 reactions
- Nov 9, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.19.0 → 11.20.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.18.0 → 11.19.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026agent-skills.github.nyldn.claude-octopus: 11.18.0 → 11.19.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.18.0 → 11.19.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026codex-plugins.github.nyldn.claude-octopus: 11.18.0 → 11.19.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026agent-skills.github.nyldn.claude-octopus: 11.17.1 → 11.18.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.17.1 → 11.18.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026codex-plugins.github.nyldn.claude-octopus: 11.17.1 → 11.18.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.17.1 → 11.18.0 — sudosubin/agents.nix0 reactions
- Nov 8, 2026Upstream sync: v11.23.0 — AlobarQuest/claude-octopus0 reactions · 1 comment
- Nov 6, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.17.0 → 11.17.1 — sudosubin/agents.nix0 reactions
- Nov 6, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.17.0 → 11.17.1 — sudosubin/agents.nix0 reactions
- Nov 6, 2026agent-skills.github.nyldn.claude-octopus: 11.17.0 → 11.17.1 — sudosubin/agents.nix0 reactions
- Nov 5, 2026codex-plugins.github.nyldn.claude-octopus: 11.17.0 → 11.17.1 — sudosubin/agents.nix0 reactions
- Nov 5, 2026agent-skills.github.nyldn.claude-octopus: 11.16.0 → 11.17.0 — sudosubin/agents.nix0 reactions
- Nov 5, 2026codex-plugins.github.nyldn.claude-octopus: 11.16.0 → 11.17.0 — sudosubin/agents.nix0 reactions
- Nov 5, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.16.0 → 11.17.0 — sudosubin/agents.nix0 reactions
- Nov 5, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.16.0 → 11.17.0 — sudosubin/agents.nix0 reactions
- Nov 3, 2026claude-code-marketplaces.github.nyldn.claude-octopus: 11.15.4 → 11.16.0 — sudosubin/agents.nix0 reactions
- Nov 3, 2026codex-plugins.github.nyldn.claude-octopus: 11.15.4 → 11.16.0 — sudosubin/agents.nix0 reactions
- Nov 3, 2026claude-code-plugins.github.nyldn.claude-octopus: 11.15.4 → 11.16.0 — sudosubin/agents.nix0 reactions
- Nov 3, 2026agent-skills.github.nyldn.claude-octopus: 11.15.4 → 11.16.0 — sudosubin/agents.nix0 reactions
- Nov 2, 2026Upstream sync: v11.15.4 — AlobarQuest/claude-octopus0 reactions · 1 comment
- Nov 1, 2026agent-skills.github.nyldn.claude-octopus: 11.15.3 → 11.15.4 — sudosubin/agents.nix0 reactions
- Oct 8, 2026feat(agent-engineering): piloto isolado dos quatro padrões com gates e evidências — Videirafo/SaaS-Engineering-Playbook0 reactions · 2 comments
- Aug 19, 2026[WR] implement claude octopus as a github action -- — midnghtsapphire/revvel-standards0 reactions · 27 comments
On Bluesky · 2
- Mar 31, 2026Since individual models have their blind spots, ensembling different coding agents can lead to more robust results. See e.g. Claude… — @andybrnr.bsky.social2 likes and reposts
- Jun 11, 2026https://github.com/nyldn/claude-octopus — @tom-doerr.bsky.social0 likes and reposts · 1 reply
On Hacker News · 1
- Mar 18, 2026Multi-tentacled orchestrator for Claude Code1 point
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- DentalEdge-Solutions/claude_code★ 1same name
- maxenko/claude-skills★ 1same name
- windgeek/skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository125 other files it gives its agents.
- AGENTS.mdAGENTS.md
- docs/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- skill-extractSkill
- flow-defineSkill
- flow-deliverSkill
- flow-developSkill
- flow-discoverSkill
- flow-parallelSkill
- flow-specSkill
- skill-agent-topologySkill
- octopus-architectureSkill
- skill-auditSkill
- skill-authoringSkill
- skill-code-reviewSkill
- skill-content-pipelineSkill
- skill-context-detectionSkill
- skill-copilot-providerSkill
- skill-cost-projectionsSkill
- skill-coverage-auditSkill
- skill-debateSkill
- skill-debugSkill
- skill-decision-supportSkill
- skill-deckSkill
- octopus-researchSkill
- skill-design-lineageSkill
- skill-doc-deliverySkill
- skill-doc-syncSkill
- skill-doctorSkill
- skill-factorySkill
- skill-finish-branchSkill
- skill-intakeSkill
- skill-intent-contractSkill
- skill-issuesSkill
- skill-iterative-loopSkill
- skill-knowledge-workSkill
- skill-meta-promptSkill
- skill-native-escalation-routingSkill
- skill-parallel-agentsSkill
- skill-prdSkill
- skill-pressure-testSkill
- skill-prototypeSkill
- octopus-quickSkill
- skill-resumeSkill
- skill-review-responseSkill
- skill-rollbackSkill
- octopus-security-auditSkill
- skill-security-framingSkill
- skill-shipSkill
- skill-staged-reviewSkill
- skill-statusSkill
- skill-task-managementSkill
- skill-tddSkill
- skill-thought-partnerSkill
- octopus-ui-ux-designSkill
- skill-verification-gateSkill
- skill-visual-feedbackSkill
- skill-work-slicingSkill
- skill-writing-plansSkill
- The first 60 shown.
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 9 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈9
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 11
Never role-play other models or invent their scores.
- Distinctive: line 33
Keep this absolute root in the workflow context and include its assignment in each new shell invocation.
- Distinctive: line 35
If the judge adapter or score validator is missing at this root, stop and name the missing component rather than running a command with an …
- Distinctive: line 40
Read the supplied artifact and relevant project instructions.
- Distinctive: line 46
Before any external provider call, including model-assisted research, inspect the candidate, supporting evidence, and accessible project …
- Distinctive: line 63
Ask only for missing information that prevents a useful, bounded run.
- Very specific: line 75
Create a unique run directory in the user's project under .claude-octopus/refine/<run-id>/, never in the installed plugin.
- Very specific: line 79
Keep the original, best candidate, rubric, model responses, score reports, and an append-only iterations.jsonl owned by …
- Very specific: line 93
Use the host's native shell, file inspection, and task tools.
- Distinctive: line 97
Before each phase, read its entire resource from the resolved installed plugin root, using the exact phase paths listed below.
- Distinctive: line 108
Before external research or judging, inspect candidate, evidence and accessible context for secrets as required above.
- Very specific: line 123
Always use refinement-pair.py through its mandatory single-judge adapter.
- Distinctive: line 136
Stop after two stalled rounds or any budget, evidence, cancellation or provider blocker.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Rubric-guided refinement
Execute the workflow below when invoked through /octo:refine or
/octo:ladder. This is a host-managed workflow using real provider calls.
Never role-play other models or invent their scores. The score validator checks
structure and arithmetic; it does not verify the truth of a model's judgments.
Set OCTO_SKILL_FILE to the absolute path of this loaded SKILL.md supplied by
the host. Resolve its installation before reading resources or running commands.
This works without a stable plugin link or trusted startup hooks. If the host
supplies no skill path, use its explicit plugin root; stop if neither resolves.
OCTO_ROOT="${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${OCTO_ROOT:-}}}"
if [[ -n "${OCTO_SKILL_FILE:-}" ]]; then
[[ "$OCTO_SKILL_FILE" == /* && -f "$OCTO_SKILL_FILE" ]] || exit 1
OCTO_ROOT="$(cd "$(dirname "$OCTO_SKILL_FILE")" && pwd -P)"
while [[ "$OCTO_ROOT" != / && ! -f "$OCTO_ROOT/.claude-plugin/plugin.json" && ! -f "$OCTO_ROOT/.codex-plugin/plugin.json" ]]; do
OCTO_ROOT="$(dirname "$OCTO_ROOT")"
done
fi
[[ -n "$OCTO_ROOT" && -f "$OCTO_ROOT/scripts/refinement-judge.sh" && -f "$OCTO_ROOT/scripts/refinement-score.py" && -f "$OCTO_ROOT/scripts/refinement-run.py" && -f "$OCTO_ROOT/skills/skill-refine/SKILL.md" ]] || {
printf '%s\n' 'Cannot resolve this installed Octopus skill' >&2
exit 1
}Keep this absolute root in the workflow context and include its assignment in each new shell invocation. Shell variables may not persist between tool calls. If the judge adapter or score validator is missing at this root, stop and name the missing component rather than running a command with an empty root.
1. Establish scope and budget
Read the supplied artifact and relevant project instructions. Accept text, files, plans, code, images, or other artifacts the host can actually inspect. For binary, visual, or remote artifacts, obtain a faithful readable or rendered snapshot with the appropriate host tools. If you cannot inspect it, stop and name the missing capability. Do not score an unseen attachment or a URL title.
Before any external provider call, including model-assisted research, inspect the candidate, supporting evidence, and accessible project context for credentials and private tokens. Repeat this inspection for new content before dispatch. If secrets are found, stop the run before research or judging and report the affected input or file without quoting its secret values. Do not continue by creating a sanitized copy, changing working directories, or asking a judge to ignore the original. The current adapter restricts writes and tools; it does not enforce isolation from the secret-bearing original or other readable files. Secret-bearing inputs require an execution boundary that prevents those reads, which this adapter does not provide. Host inspection is not an automatic secret scanner or a guarantee of data isolation.
Record the audience, intended outcome, constraints, editable paths, and what must be preserved. A request to improve a plan authorizes editing the plan, not executing the project described by it. A scoring or review-only request produces assessments and proposed revisions without changing the original. Otherwise invocation authorizes scoped improvements to the supplied artifact. Ask only for missing information that prevents a useful, bounded run.
Validate all options before work. Defaults are target 95/100, max five
improvement rounds, max 30 elapsed minutes, and max 14 external model calls.
--max-rounds accepts integers 1 through 10; minutes and calls must be positive
integers; target must be finite and between 0 and 100. Preserve explicit budgets.
Retries, research, scoring, and model-assisted edits all consume the call budget.
Baseline is round zero. Check remaining time and calls before every dispatch,
and reserve two scoring calls before starting an improvement. Bound each call
by the remaining elapsed budget. Never reset counters on resume or restart a
run to evade a limit. Stop after two rounds without verified improvement.
Create a unique run directory in the user's project under
.claude-octopus/refine/<run-id>/, never in the installed plugin. Check the
project's ignore rules before writing sensitive content there. If it is not
ignored, use a private directory outside the repository and report its path.
Keep the original, best candidate, rubric, model responses, score reports, and
an append-only iterations.jsonl owned by scripts/refinement-run.py. Never
append hand-written iteration decisions. Record hashes, actual model IDs and effort,
calls, elapsed time, available usage/cost data, changes, and reasons for keeping
or rejecting candidates. Unknown monetary cost stays unknown.
Snapshot the exact initial content and existing dirty edits. Work on a candidate
copy. Publish only an accepted candidate into the authorized paths after checking
that the original has not changed since it was read. Preserve concurrent edits
and stop on a conflict. Never use git reset --hard, git clean, blanket
checkout, git add -A, or automatic commits to manage these experiments.
Phase loading and fixed judging contract
Use the host's native shell, file inspection, and task tools. On Codex, load
skills/blocks/codex-host-adapter.md from the resolved plugin root. Do not assume
that a host exposes a provider or subagent until its tools confirm availability.
Before each phase, read its entire resource from the resolved installed plugin root, using the exact phase paths listed below. These paths work for both Claude and Codex installations. Stop if a resource is missing. Load only the next phase, and reload its resource after compaction before resuming it:
- Scope, secret inspection, candidate preservation, and budgets are above.
- Load
skills/skill-refine/references/phase-2.mdto select and freeze the rubric. - Load
skills/skill-refine/references/phase-3.mdbefore independent scoring. - Load
skills/skill-refine/references/phase-4.mdbefore edits and rescoring. - Load
skills/skill-refine/references/phase-5.mdbefore delivery and trajectory.
Before external research or judging, inspect candidate, evidence and accessible
context for secrets as required above. Run declared guards before model review.
Choose exactly Claude and Codex with preflight.sh --providers claude,codex,
without changing the council roster. Preserve fixed explicit models and efforts,
record requested and resolved identities, and stop on a required provider failure.
A readiness receipt is selection evidence only. Fresh dispatch auth, quota, and
policy checks remain mandatory, including when credentials change after selection.
Both independent read-only judges must receive the same frozen candidate, blind
rubric, and guard evidence. Hide target, prior scores, author and other judges.
Initialize the managed run through refinement-run.py as described in phase 2.
Its begin, prepare, and record commands enforce round admission, guard
failures, acceptance and two-round stalls. A stopped run cannot be reinitialized
or moved to another directory to reset its limits. Keep legacy runs read-only
until their evidence has been reconciled; never invent controller state for them.
Always use refinement-pair.py through its mandatory single-judge adapter.
Reserve both calls before launch and count failures, cancellation, research,
retries, and resume against the same total budget and absolute deadline. Never
redispatch a reserved pair ID. Raw stdout and one successful seat cannot pass.
For a material rubric correction, follow phase 2's explicit version transition
and rescore the original before the best candidate within the remaining budget.
Bind assessment identities to actual dispatch metadata, exactly one canonical
claude and one codex, plus rubric and candidate hashes. Run
scripts/refinement-score.py on both valid assessments. Its conservative lower
score, criterion minima, no blockers, passing guards and disagreement limit of
10 points control acceptance. Verify findings before edits. Never average away
failure or weaken the rubric. Accept only verified improvement with no regression.
Stop after two stalled rounds or any budget, evidence, cancellation or provider
blocker. Resume only with matching ledger, original, candidate, current rubric version, fixed
panel, models and efforts; never reset counters. Label target success only when
the validator reports target_met: true for the delivered bytes. Preserve
concurrent edits and existing approval boundaries through delivery.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

