OpenCoven Role Creation Process
Skill75 starsChanged 4 months agoMITView on GitHub
Create, validate, and activate ROLE.md specs for familiars with low-lift agentic guidance.
In this repository59 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- skills/heygen-skills/CLAUDE.mdCLAUDE.md
- 1passwordSkill
- ask-curlSkill
- boardSkill
- cleanup-unused-filesSkill
- codeflow-maintainerSkill
- codex-session-managerSkill
- coven-astraSkill
- coven-board-entrySkill
- coven-charmSkill
- coven-cloud-maintainerSkill
- coven-codySkill
- coven-echoSkill
- coven-identity-canonSkill
- coven-kittySkill
- coven-parallel-workSkill
- coven-sageSkill
- coven-task-managerSkill
- daily-dev-agenticSkill
- discrawlSkill
- disk-space-optimizerSkill
- epubSkill
- familiar-board-stewardshipSkill
- heygen-avatarSkill
- heygen-videoSkill
- heygen-skillsSkill
- higgsfield-generateSkill
- higgsfieldSkill
- linear-issue-managementSkill
- lit-ui-designerSkill
- lobsterSkill
- maintainer-barSkill
- memory-recallSkill
- memory-timeline-managerSkill
- ml-engineerSkill
- nodriverSkill
- ocrSkill
- openclaw-devSkill
- openclaw-trustSkill
- opencoven-designSkill
- pr-agentSkill
- project-managerSkill
- prompt-engineerSkill
- prompt-vaultSkill
- research-ingestionSkill
- ritual-dapp-frontendSkill
- rollcallSkill
- security-agentSkill
- skill-scannerSkill
- slidev-deck-pipelineSkill
- tauri-apple-releaseSkill
- threejs-animationSkill
- tinyfish-agent-runSkill
- tinyfish-browserSkill
- tinyfish-fetchSkill
- tinyfish-searchSkill
- tinyfishSkill
The file
We split this skill into the rules it gives an agent
12 rules. About 0 are common, 8 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
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 15
Use this skill when a CovenCave user wants to create, review, revise, or standardize Roles for familiars.
- Distinctive: line 23
If the familiar has no workspace yet, create or select one before writing Roles.
- Distinctive: line 26
Use a Role for a domain-specific frame plus workflows and permissions.
- Distinctive: line 27
Use a Skill for one reusable procedure.
- Distinctive: line 28
Use a Workflow for one multi-step path inside an existing Role.
- Distinctive: line 29
Use a Plugin/tool grant for capability without identity context.
- Very specific: line 30
Leave core identity in SOUL.md.
- Very specific: line 33
Use templates/ROLE.md as the starting point.
- Distinctive: line 34
Keep the body specific enough to change behavior.
- Very specific: line 35
Do not duplicate the familiar's SOUL.md.
- Very specific: line 38
Every item listed in the Role's workflows: frontmatter should have a matching file under workflows/<workflow-id>.md.
- Distinctive: line 95
Keep Kitty-like generalists role-light unless a role adds concrete behavior.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
OpenCoven Role Creation Process
Use this skill when a CovenCave user wants to create, review, revise, or standardize Roles for familiars.
A Role is not a plugin list. A Role is the familiar's domain-specific way of being while doing a class of work. It bundles identity context, skills, tools, workflows, plugins, and permission declarations.
Low-Lift Agentic Flow
- Read the familiar.
- Load the familiar's
SOUL.md,IDENTITY.md, local skill doctrine, existing Roles, and any user-provided purpose. - If the familiar has no workspace yet, create or select one before writing Roles.
- Load the familiar's
- Decide whether a Role is actually needed.
- Use a Role for a domain-specific frame plus workflows and permissions.
- Use a Skill for one reusable procedure.
- Use a Workflow for one multi-step path inside an existing Role.
- Use a Plugin/tool grant for capability without identity context.
- Leave core identity in
SOUL.md.
- Draft the Role.
- Use
templates/ROLE.mdas the starting point. - Keep the body specific enough to change behavior.
- Do not duplicate the familiar's
SOUL.md.
- Use
- Draft workflows.
- Every item listed in the Role's
workflows:frontmatter should have a matching file underworkflows/<workflow-id>.md. - Use
templates/workflow.md.
- Every item listed in the Role's
- Choose canonical placement.
- Familiar-specific source of truth:
~/.coven/roles/familiars/<familiar>/<role-id>/. - Shared global source of truth:
~/.coven/roles/global/<role-id>/. - Symlink into the harness-visible familiar workspace:
<familiar-workspace>/roles/<role-id>.
- Familiar-specific source of truth:
- Set activation deliberately.
- Primary domain Roles may start active after review.
- Narrow task Roles should start inactive.
- High-authority Roles should start inactive.
- Validate.
- Run
scripts/validate-roles.mjs. - Confirm required frontmatter, listed workflows, symlinks, and
SOUL.mdrelationship text.
- Run
- Report cleanly.
- Roles created.
- Roles activated.
- Source files read.
- Validation output.
- Open questions or permission risks.
Required Role Contract
Each ROLE.md must include:
nameidversiondescriptionfamiliarskillstoolspluginsworkflowspermissions- a body with purpose, scope, principles, workflows, permissions, handoffs, and relationship to
SOUL.md.
Anti-Patterns
Avoid:
- Role as duplicated
SOUL.md. - Role with only frontmatter and no identity context body.
- One Role per tiny task.
- Broad tool access just in case.
- Personality/vibe rules that belong in core identity.
- Assuming Role permission declarations enforce themselves.
Recommended Defaults
- Authority layer wins over Role declarations.
- Cave active state is UI/session state.
- Role manifest is intent and context.
SOUL.mdwins over every Role.- Keep Kitty-like generalists role-light unless a role adds concrete behavior.
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.

