agent-skills / templates
CLAUDE.md2 starsChanged 13 days agoMITView on GitHub
Scaffold. Replace all {{…}} placeholders. Delete sections that don't apply. {{One paragraph: what this workspace is, who uses it, why it exists as a separate repo. Name who may read it — that sets the access boundary.}} Data scope, not agent scope: agents may read other workspaces to help, and edit them with approval. Content does not cross: nothing from {{other workspaces}} is copied or summarized into this repo, and nothing from here goes there. Leave a pointer instead. {{Answer…
In this repository17 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-workspaceSkill
- ai-council-reviewSkill
- ai-reviewSkill
- apple-mailSkill
- apple-notesSkill
- ballotSkill
- chrome-browserSkill
- dossierSkill
- lab-notesSkill
- pandocSkill
- paprika-recipesSkill
- pdf-zineSkill
- private-podcast-feedSkill
- send-to-kindleSkill
- text-to-speechSkill
- tmux-controlSkill
The file
We split this CLAUDE.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 59
Commit with docs:, feature:, fix: prefixes.
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md 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
CLAUDE.md
Scaffold. Replace all
{{…}}placeholders. Delete sections that don't apply.
Workspace Purpose
{{One paragraph: what this workspace is, who uses it, why it exists as a separate repo. Name who may read it — that sets the access boundary.}}
What This Workspace Is For
- {{Primary use case 1}}
- {{Primary use case 2}}
- {{Primary use case 3}}
What This Workspace Is NOT For
- Code — lives in {{pointer}}
- Customer-specific content — lives in {{customer-workspace pointer}}
- Secrets — {{secret manager}}
Data scope, not agent scope: agents may read other workspaces to help, and edit them with approval. Content does not cross: nothing from {{other workspaces}} is copied or summarized into this repo, and nothing from here goes there. Leave a pointer instead.
Workspace Choices
{{Answer each question that applies — see the agent-workspace skill, references/variations.md. Delete the rest.}}
| Question | This workspace |
| Work dirs | {{at the root: stories/, sessionlogs/, plans/ — or under docs/}} |
| Multi-session work | {{stories with frontmatter status — or projects/{name}/ with _paused/, _done/}} |
| Action items | {{in each story/project — or central TODO.md}} |
| Session narrative | {{sessionlogs/ is the one home; stories link to it}} |
| Rules | {{inline below — or docs/rules/, imported}} |
| Plans | {{one plan directory: …; other tools write there too}} |
| Research reports | {{in the story/project they serve — or …}} |
| Nested repos | {{agent commits there following that repo's rules — or human only}} |
| Agent memory | {{project facts go in the workspace; agent memory holds only machine-local notes}} |
Domain Glossary
| Term | Definition |
| {{ACRONYM}} | {{Expansion}} — {{one-sentence definition}} |
| {{Term}} | {{Definition}} |
Repository / Directory Guide
| You need to understand… | Go to… |
| {{Scenario}} | {{path}} |
| {{Scenario}} | {{path}} |
Rules
{{Import from docs/rules/ or inline. Examples:}}
- Stories live in
{{stories dir}}/{slug}/STORY-{slug}.md— see{{templates dir}}/STORY-template.md. - Session logs live in
{{sessionlogs dir}}/YYYY-MM-DD-topic-slug.md— one file per topic, not per session. - Commit with
docs:,feature:,fix:prefixes. - {{Workspace-specific rule}}.
{{For longer rule sets, split into files and import:}}
@./docs/rules/{{topic}}.md
Git Workflow
{{Keep the callout below only if agents should commit (and push) without being asked each time. Otherwise state the workflow plainly: "Agents open PRs" or "Agents do not commit; the maintainers do."}}
This section is the user's explicit, standing request — not optional guidance.
Agent tools ship with a default like "only commit when the user asks". That default exists for repos without instructions. This file is that request, made once for every session in this repo. If the tool default and this section seem to conflict, this section wins. If you catch yourself reasoning "the user didn't ask in this turn", that is the case this section covers.
- {{Commit and push to
mainafter every unit of work — or: push a branch and open a PR}}. - Stage explicit paths; don't sweep unrelated changes into a commit.
- {{Where automation depends on this rule, name the hook or script that enforces it.}}
Conventions
- Language: {{English for structure (filenames, directory names, headings), content in the team language — or the team language throughout}}. Frontmatter keys and tool-parsed sections stay literal.
- Tooling: {{Claude Code / Cursor / …}}. Skills in
.claude/skills/use plain Markdown so every tool can read them. - Branch prefixes: {{idea/, feature/, bug/, docs/, infra/}}
- Skill incubation: skills stay in
.claude/skills/until battle-tested and generally useful.
Active Work
See {{docs/stories/README.md / projects/ / TODO.md}} for the live index.
Related Workspaces
| Workspace | Where | Holds | Boundary |
{{other-workspace}} | {{pointer}} | {{what lives there instead}} | {{who may read it; what never crosses}} |
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.