record your first skill

intent

convert a tacit workflow into a structured, conditional, measurable skill via demonstration. take one workflow you already do by hand every week, demonstrate it once, answer a few questions, and now anyone in your org can invoke it by name. the interview step is what makes it a skill instead of a transcript. you are not just recording a recipe, you are teaching the model when to fire it, what to do if a step fails, and what success looks like. trigger this skill when the user says "i do X every week" or "i wish claude remembered how to do this".

inputs

• one workflow you do repeatedly (weekly is the sweet spot, daily is also solid)
• the tools you normally use, already authenticated (mcp servers, slack, gmail, github, salesforce, hubspot, etc.)
• 10-15 minutes uninterrupted to demo plus answer questions
• stable network connection (recordings pause if connection drops; resume when restored)
• access to the /implexa:record-skill command and org-level skill creation permissions

procedure

1. pick the right workflow

   • input: your mental list of recurring tasks
   • output: one workflow selected
   • candidates that work well: weekly status email (pull commits, format, send), ticket triage (check duplicates, create issue, comment), lookup-then-summarize (account research, candidate sourcing, contract review), scheduled report (daily standup, weekly churn scan, monthly board update)
   • candidates that do not work well: one-off creative work (write the launch tweet), anything with truly novel decisions every time, workflows where the steps change weekly
   • guidance: if the workflow takes more than 15 minutes by hand, split it. shorter skills compose better than mega-skills.

2. start recording

   • input: the selected workflow, all required tool access
   • output: recording indicator live, demonstration log initialized
   • action: call /implexa:record-skill
   • guidance: narrate decisions out loud as you work ("checking email first because it has the customer id"). every tool call, every decision, every output gets captured automatically.

3. perform the workflow end-to-end

   • input: the live demonstration context, access to all tools
   • output: complete demonstration trace with all steps, tool calls, and decision points captured
   • action: execute the workflow as you normally would, speaking aloud about why you are doing each step
   • guidance: work at your normal pace. if you make a mistake mid-demo, keep going and mention it in free-text addendum later. the interview will route around it.

4. end recording

   • input: completed workflow demonstration
   • output: demonstration capture finalized, ready for interview
   • action: call end_demonstration() or say "done"
   • guidance: the live capture stops. the system now has a full trace of what you did.

5. add optional free-text notes

   • input: the demonstration trace, any edge cases or branching logic you did not explicitly show
   • output: freetext addendum appended to demonstration log
   • action: call record_demo_freetext() with any additional context
   • guidance: mention edge cases ("if X happens do Y"), alternative branches you did not demo, or tool calls you want stripped from the final skill (api keys, personal accounts, etc.)

6. answer the interview

   • input: the demonstration trace and your freetext notes
   • output: interview answers, decision point definitions, trigger phrases, success criteria
   • action: a haiku-class model asks 3-8 questions tuned to what it saw you do
   • typical questions: what should this skill be named? when should it fire (trigger phrases)? what counts as a successful outcome? what should it do if [tool] returns an error? is there a step where the user should confirm before proceeding?
   • guidance: answer briefly. the model needs signal, not essays. if you do not have an answer, say "skip" or "leave it open". the agent picks a sensible default. you can refine via /implexa:update-skill later.

7. finalize and save

   • input: the complete demonstration, freetext notes, interview answers
   • output: a saved org skill with slug, name, description, 6-component body, trigger phrases, demonstration trace, interview answers
   • action: the agent assembles a 6-component SKILL.md, scrubs anything that looks like a secret or pii, and saves it scoped to your org
   • guidance: you get a confirmation with the slug. anyone in your org can now invoke it by name.

decision points

• the interview asks a question you do not have an answer for: say "skip" or "leave it open". the agent picks a sensible default. you can refine via /implexa:update-skill later.

• the demonstration captured a tool call you do not want in the final skill (api key, personal account, internal id): mention it in the free-text addendum or strip it via /implexa:update-skill after save. the system flags pii automatically but manual review is wise.

• the workflow has a branch you did not demo ("if X, do Y; if Z, do W"): demo the most common path, then describe the other branches in free-text. the agent encodes them as explicit decision points in the final SKILL.md.

• recording was interrupted (network drop, timeout): the system pauses the recording. resume when connection is restored. the demonstration log stays intact up to the interruption point.

• the selected workflow is too long (takes more than 20 minutes): split it into two smaller skills. each skill should be independently invocable and compose cleanly with others.

output contract

a saved org skill visible in the org skill registry with:

• a unique slug (auto-generated from the name you provided)
• the skill name and one-line description
• the full 6-component body (intent, inputs, procedure, decision points, output contract, outcome signal)
• recorded trigger phrases (when the user says X, invoke this skill)
• the demonstration trace (readable log of tool calls and outputs, scrubbed of pii)
• the interview answers
• org-level access permissions (visible to everyone in your org immediately)
• shareable cross-org via /implexa:share-this if needed

location: available at /implexa:skill/<slug> and in org skill search results.

outcome signal

the skill gets used. if the same workflow gets invoked at least once in the 30 days after recording (by you or anyone on your team), it stuck. if it never gets used, the workflow was too one-off or the trigger phrases did not match how people actually search. check usage via /implexa:skill-usage <slug>.