engineer-plan
Skill2 starsChanged 2 months agoMITView on GitHub
Turn context.md and architecture.md into a phased plan.md
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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 14
Read context.md and architecture.md in .codex/work/<type>/<slug> if you haven't already.
- Distinctive: line 25
If you're working on this, keep this plan.md file updated as you progress.
- Distinctive: line 67
Use file search/grep tools to find specific files based on your discovery answers, and read through relevant code in batches.
- Distinctive: line 69
Use web search for best practices or library documentation when needed.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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
Engineer Plan
This is the skill to start planning a feature's execution.
The user's request should name the work item as <type>/<slug> (e.g. feat/csv-order-export), matching the work item created by engineer-context and engineer-architecture. If they didn't state it clearly, ask before proceeding rather than guessing.
Review
Read context.md and architecture.md in .codex/work/<type>/<slug> if you haven't already. If either file doesn't exist yet, stop and tell the human to run engineer-context and/or engineer-architecture first — don't invent a plan without them.
Your task now is to create a detailed execution plan, saved to .codex/work/<type>/<slug>/plan.md. The goal of this document is a phased execution approach that lets us build the feature incrementally, testing each phase as we go, and that makes it possible to resume work if you get interrupted.
plan.md should split execution into phases, each phase sized to a chunk of work a human could complete in about 2 hours.
Template for plan.md:
<plan>
# [WORK NAME]If you're working on this, keep this plan.md file updated as you progress.
PHASE 1 [Done ✅]
Details of this part of the feature.
A task that was done [Done ✅]
Details about the task.
A task that was done [Done ✅]
Details about the task.
Notes:
- Something that happened and forced a change of direction.
- Something we learned while building.
- Something we discussed and agreed on.
PHASE 2 [In Progress ⏰]
A task that needs doing [In Progress ⏰]
Details about the task.
A task that needs doing [Not Started ⏳]
Details about the task.
PHASE 3 [Not Started ⏳]
A task that needs doing [Not Started ⏳]
Details about the task.
A task that needs doing [Not Started ⏳]
Details about the task.
</plan>Tips:
- Use file search/grep tools to find specific files based on your discovery answers, and read through relevant code in batches.
- Examine specific implementation details.
- Use web search for best practices or library documentation when needed.
If this research surfaces a new architectural decision or a contradiction with earlier decisions, start a discussion about it with the human, agree on the changes, and update architecture.md for this feature if needed. If the decision is significant enough to constrain future work beyond this feature, suggest recording it with engineer-adr.
This artifact should also note which tasks need to happen sequentially versus which can run in parallel.
Once plan.md is finalized, tell the human you're ready to move to the next step.
⛔ Don't proceed automatically. Wait for the human to approve the plan and explicitly run engineer-work. Don't start implementation on your own.
In this repository31 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- bootstrap-business-docsSkill
- bootstrap-indexSkill
- bootstrap-tech-docsSkill
- engineer-adrSkill
- engineer-architectureSkill
- engineer-bumpSkill
- engineer-contextSkill
- engineer-coverageSkill
- engineer-discoverSkill
- engineer-doctorSkill
- engineer-pre-prSkill
- engineer-prSkill
- engineer-reviewSkill
- engineer-sync-docsSkill
- engineer-validateSkill
- engineer-workSkill
- meta-create-agentSkill
- meta-preferencesSkill
- product-brainstormSkill
- product-collectSkill
- product-quick-specSkill
- product-refineSkill
- product-specSkill
- product-sync-githubSkill
- product-validateSkill
- python-developerSkill
- typescript-developerSkill
- warm-upSkill
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.