forgetful-procedures
Skill341 starsChanged 23 days agoMITView on GitHub
Store and find procedural knowledge — reusable step-by-step how-tos — in Forgetful's skill store. Use when a workflow worth repeating gets worked out, when the user asks how something is done here, or when importing or exporting SKILL.md files. Litmus: a procedure someone follows is a skill; a fact someone recalls is a memory.
In this repository11 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- forgetful-cli-setupSkill
- forgetful-context-gatherSkill
- forgetful-encode-repoSkill
- forgetful-entitiesSkill
- forgetful-exploreSkill
- forgetful-filesSkill
- forgetful-mcp-setupSkill
- forgetful-recallSkill
- forgetful-rememberSkill
The file
We split this skill 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 85
Use import_skill_bundle / export_skill_bundle when references, scripts or binary assets must travel with SKILL.md.
≈ “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
Storing and finding procedures
Forgetful's skill store is procedural memory: how-tos that agents retrieve by capability and follow. All skill operations run through the standard invocation surface below, the same as every other domain.
Invoking operations
Operations are named by registry name (search_skills, create_skill, ...). Invoke via
whichever surface this agent has:
- MCP:
execute_forgetful_tool(tool_name="search_skills", arguments={...}) - CLI:
forgetful call search_skills --args '{"query": "..."}' --json
Get any operation's schema at runtime: how_to_use_forgetful_tool (MCP) or
forgetful tools info <operation> (CLI) — schemas are deliberately not repeated here.
Step 1 — Apply the litmus
"How to review pull requests in this project" is a skill; "we use conventional commits" is
a memory (forgetful-remember). A procedure has steps someone follows in order; a fact is
recalled and applied.
Done when: the content is confirmed procedural.
Step 2 — Search before create
search_skills for the capability first — search is semantic over skill descriptions, so
query by what the procedure achieves ("deploy the API", "rotate credentials"), not its
exact name. An existing skill that covers it gets updated (update_skill) rather than
duplicated; names are unique.
Done when: create vs update is an informed decision.
Step 3 — Create with the description carrying the weight
The description is the only embedded field — content (up to 100KB) is stored but never
searched semantically. Write the description as the retrieval surface: what the procedure
achieves, when to reach for it, capability keywords. Then create_skill with kebab-case
name, the full markdown procedure as content, tags, and importance.
Done when: the skill exists and a capability-phrased search_skills finds it.
Step 4 — Import / export SKILL.md
import_skill ingests a SKILL.md string: YAML frontmatter needs name and description;
license, compatibility, tags, metadata, and allowed-tools (hyphenated on disk,
mapped to allowed_tools internally) are honoured. The body after the frontmatter becomes
the content. YAML references (*name referring to an &name anchor) are rejected in
frontmatter; write the values directly instead. Quoted "*person" and body asterisks are valid.
export_skill re-emits the SKILL.md, round-trip faithful — tags included.
Done when: imported skills search well (check the description survived intact) and exports open cleanly in the target platform.
Step 5 — Link the supporting material
Procedures rarely stand alone: link_skill_to_memory for the decisions behind the
procedure, link_skill_to_document for long-form background, link_skill_to_code_artifact
for the code it operates, link_skill_to_file for bundled assets. get_skill_links shows
everything attached — it is the reverse lookup, returning linked memory, file, code
artifact, and document IDs.
Done when: following the procedure never requires hunting for its supporting material.
Portable supporting files
Use import_skill_bundle / export_skill_bundle when references, scripts or binary assets must travel with SKILL.md. Discover their schemas; provide each resource with its package-relative path and required fields. Existing Markdown-only operations remain unchanged. See docs/skill-bundles.md for canonical text, limits, permissions and the CLI argument-size restriction.
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.

