promote-to-tooling
Skill6 starsChanged 29 days agoMITView on GitHub
Review a tooling change made in a working repo (new skill, script, rule, hook, template, or workflow fix) and promote the universal parts into the clybor-claude-tooling master template. Use when the user says "promote to tooling", "add this to the master tooling", "update clybor-claude-tooling", "parlay this into the template", "should this go in the template repo", or right after building or adapting tooling in any project repo. Runs a universality test, diffs against the master counterpart, generalizes project-specific bits, applies the edit, wires references, and flags downstream repos that need a re-sync.
In this repository30 other files it gives its agents.
- adversarial-re-readSkill
- adversarial-reviewSkill
- build-estimateSkill
- build-evaluateSkill
- build-executeSkill
- build-planSkill
- build-polishSkill
- build-prdSkill
- build-probeSkill
- build-validateSkill
- excel-connector-governanceSkill
- five-whysSkill
- insight-builderSkill
- insight-crystallizerSkill
- insight-promotionSkill
- multi-sessionSkill
- notion-governanceSkill
- oodaSkill
- prd-writerSkill
- quality-reviewSkill
- ralph-implementSkill
- ralph-loopSkill
- search-governanceSkill
- session-outputSkill
- skill-validatorSkill
- task-planSkill
- verifySkill
- voice-profile-testSkill
- why-diagnosticSkill
- writing-qualitySkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 16
When the user asks whether a change belongs in the master
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
Promote to Tooling
clybor-claude-tooling (~/gits/clybor-claude-tooling/) is the master template new repos
initialize from (scripts/init.sh). Tooling improvements born in working repos
(life-crm, client repos) die there unless promoted. This skill is the promotion path.
When to run
- Right after building or meaningfully adapting a skill, script, rule, hook, command, agent, or template in a working repo
- When the user asks whether a change belongs in the master
- Periodically: "review recent tooling changes for promotion candidates"
The universality test (gate)
From the master rules README: would this apply equally to a Python backend, a TypeScript frontend, a research repo, a documentation site?
| Verdict | Action |
| Universal as-is | Port verbatim |
| Universal mechanism, project-specific surface (paths, names, IDs) | Generalize, then port |
| Project-specific (registry formats, client branding, Notion IDs, MCP tool IDs) | Stays in the working repo — record a pointer only |
Where things live in the master
| Working-repo artifact | Master destination |
| Skill | .claude/skills/<name>/SKILL.md |
| Slash command | .claude/commands/<name>.md |
| Agent definition | .claude/agents/<lane>/<name>.md |
| Hook script | .claude/hooks/ + wire in .claude/settings.json.template |
| Always-on rule | .claude/rules/ (only if universal — see rules README) |
| Gate/utility script | scripts/ |
| Document/folder template | templates/ |
| Router or Hard-Rule text | templates/CLAUDE.md.template |
Workflow
- Name the candidate — file path(s) in the working repo + one line on what it does.
- Universality test — verdict from the table above. Stop here if project-specific.
- Diff against the master counterpart — search the master for an existing equivalent
(skills list,
scripts/,templates/, commands). If one exists, diff and merge into it — don't create a parallel duplicate. - Generalize — strip project paths, person names, repo-specific IDs. Pick the right default over adding a config (KISS).
- Apply — in Cowork,
.claude/paths need Desktop Commander on the host path (~/gits/clybor-claude-tooling/...);scripts/,templates/,docs/take native writes. In Claude Code, edit directly. - Wire the references — router row and Key Paths in
templates/CLAUDE.md.template, skills-list row,settings.json.templatehook if applicable. Undiscoverable tooling is dead tooling. - Validate — run a changed script once; run
bash scripts/check-context-budget.shif always-on files grew; confirm new skill frontmatter parses (name + description, description under 1024 chars). - Report downstream impact — repos already initialized do NOT auto-update. List
which active repos should re-sync the changed file (manual copy, or re-run the
relevant
init.shstep). The working repo that originated the change usually keeps its own specialized version — say so explicitly.
Anti-patterns
- Porting registry- or client-coupled scripts wholesale (life-crm's
new-project.shis registry-specific; the master's analog isinit.sh) - Creating a parallel skill when the master already covers the concept under another
name (e.g.,
prd-writer/task-planvs a working repo'sbuild-prd/build-plan— diff and merge, don't fork) - Promoting a non-universal rule into
.claude/rules/— rules are paid for in always-on context in every downstream repo - Promoting without updating
templates/CLAUDE.md.template— the next repo won't know the tooling exists - Skipping the downstream-impact report — a master-only fix silently diverges from every live repo
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.

