conon.ai

conon.ai/Tips

Typed It Three Times? Make A Skill

The third time you type the same procedure is the signal: say 'create a skill that does this' and the agent writes the skill file itself. Then improve it in a loop — run it, note what's off, have it edit its own skill. House rules: 2nd repeated mistake → CLAUDE.md line; 3rd repeated prompt → skill.

AdvancedAUTOMATIONClaude Code

Why it works

The third time you type the same procedure is the signal — not the first (novelty) or second (coincidence). Say 'create a skill that does this' and the agent writes its own skill file; from then on the procedure is one command. Then improve it like a process, not a prompt: run, critique, have it edit its own skill. House rules keep the artifacts sorted: repeated MISTAKE → CLAUDE.md line; repeated PROMPT → skill.

How to do it

  1. Notice the third repetition of a procedure you're typing out.
  2. Tell the agent: 'create a skill that [the procedure]' — it writes the skill file itself.
  3. Run the skill on a real case immediately.
  4. Critique the output; then: 'update your skill file so that next time [the gap] is handled.'
  5. Repeat the run-critique-patch loop 2–3 times; the skill converges fast.
  6. Invoke by command or just describe the task — the agent picks up its own skills contextually.

Copy this

prompt
Create a skill that prepares a Telegram post from my notes using my template: hook line, three bullets, one CTA, no hashtags. Write the skill file, then run it on the notes below as a test.

[paste notes]

What to watch for

  • Skills hold PROCEDURES; CLAUDE.md holds RULES. Stuffing procedure into the briefing taxes every session; stuffing rules into skills means they're only followed when the skill runs.
  • A skill you never iterate is a prompt with extra steps — the second draft is where skills start beating retyping.

Sources

Original: conon.aiDocumentation: code.claude.com

Read next

Scheduled Agents Need An Open Lid
Local scheduled tasks run only while the app is open and the machine is awake — a missed slot is skipped, not postponed. Turn on keep-awake, leave the laptop lid open, and write drift-tolerant instructions: 'if it's already past 5pm, skip today and report that you skipped'. Need lid-closed reliability? That's a cloud task — which can't see your local files.
Claude Code
Writer Session, Reviewer Session
Run two parallel sessions on one version-controlled project: one builds, the other — with fresh context — reviews. The docs are explicit about why it works: a fresh session isn't attached to code it didn't write, so it critiques honestly where the author session defends.
Claude Code
Send Noisy Research To A Subagent
Any side-investigation that would flood your session — 'which of these fifty files mentions the old pricing?' — goes to a subagent. It reads everything inside its own separate memory and returns three lines; your session's context stays clean for the actual work.
Claude Code
/context Names What Eats The Window
When the agent 'gets dumber toward evening', the window is full — /context shows exactly what by: tools, skills, messages, briefing. Then choose deliberately: /compact WITH instructions on what to preserve (it's an expensive operation — the agent re-reads everything), or the free /clear if the task is done.
Claude Code

We build and operate the systems behind tips like this one — agents, automations and internal tools, shipped into production. See what the studio does →