automation-pick
Skill1 starsChanged 23 days agoMITView on GitHub
Before automating a task, decide whether to automate. Decision tree based
In this repository46 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .codex/AGENTS.mdAGENTS.md
- .gemini/AGENTS.mdAGENTS.md
- .cursor/rules/agent-foundry-orchestrator.mdcCursor rule
- .cursor/rules/agent-foundry-skills.mdcCursor rule
- web/llms.txtllms.txt
- anti-slopSkill
- api-designSkill
- bottleneck-gatingSkill
- budget-guardSkill
- cache-identitySkill
- constraint-then-solveSkill
- context-optimizationSkill
- cron-troubleshootSkill
- e2e-test-strategySkill
- engram-routingSkill
- execution-observabilitySkill
- fallback-model-routingSkill
- feedback-loopSkill
- judge-integritySkill
- knowledge-extractSkill
- landscape-firstSkill
- measure-firstSkill
- plan-before-codeSkill
- plan-then-actSkill
- prompt-disciplineSkill
- pushback-when-wrongSkill
- quality-protocolSkill
- read-before-buildSkill
- re-verify-findingsSkill
- session-closeoutSkill
- session-distillSkill
- show-your-workSkill
- structured-output-contractSkill
- verify-firstSkill
- workflow-decomposeSkill
- agent-memory-vaultSkill
- agent-team-standardsSkill
- autonomous-briefingSkill
- book-to-skillSkill
- chrome-devtools-mcp-bridgeSkill
- design-languageSkill
- funnel-pr-guardSkill
- persistent-memorySkill
- sql-migration-trioSkill
- token-compressionSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 33
If no → don't automate yet.
- Distinctive: line 37
If no → don't automate. Even a 2-hour script doesn't pay back.
- Distinctive: line 50
If the rules keep changing, or humans make judgement calls → don't automate (yet).
- Distinctive: line 54
If the automation runs and gets the wrong answer:
≈ “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
Automation Pick
A framework for deciding which tasks are worth automating. Most tasks aren't.
When to Use
- Someone says "we should automate this"
- You're tempted to write a script for a one-off task
- Reviewing a backlog of "automate X" requests
- Building automation that turned out to be wrong
Procedure
The Decision Tree
Answer in order. Stop at the first "no".
1. Has this been done manually at least 3 times?
If no → don't automate yet. Do it manually 3 times. Patterns only emerge with repetition.
2. Will it be done ≥10 more times in the next year?
If no → don't automate. Even a 2-hour script doesn't pay back.
3. Does manual execution have a real error cost?
| Cost | Examples | Automate? |
| Catastrophic | Billing, security, data loss | ✅ Yes, urgently |
| Annoying | Wrong formatting, missed deadlines | ✅ Yes, eventually |
| Cosmetic | Looks bad | ❌ Probably not |
| Zero | "It would be nice" | ❌ No |
4. Is the task fully understood and deterministic?
If the rules keep changing, or humans make judgement calls → don't automate (yet). Standardize first, automate second.
5. Is the task reversible?
If the automation runs and gets the wrong answer:
- Easy to detect and revert → automate
- Hard to detect, propagates → don't automate without safeguards
The 4-Quadrant Matrix
| Low error cost | High error cost | |
| High volume | Automate freely | Automate carefully (with monitoring) |
| Low volume | Don't automate | Don't automate (do it manually + check) |
Worked example
A rejected automation that stops at the first failed question.
Job: rename screenshots every week.
1. Manual runs: 2 -> stop here.
Decision: do not automate yet; log the next three manual runs.
Revisit when manual runs >= 3 and forecast >= 10/year.Anti-patterns
- Writing a script for a task done twice.
- Automating a judgment call with no stable rules.
- Automating an irreversible action with no safeguard.
References
- details
Verification Checklist
- Manual runs >=3 and forecast >=10/year are recorded. (step 1)
- Error cost and volume place it in the matrix. (step 3)
- The task has deterministic rules. (step 4)
- Reversibility and detection are stated. (step 5)
- A rejected automation has a reason. (step 1)
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.

