agentleFS
youcislayoucisla/Agent-Foundry/skills/core/read-before-build/SKILL.md

read-before-build

Skill1 starsChanged 23 days agoMITView on GitHub

"A plan or design document is an aspiration, not a specification. Before

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
  1. Distinctive:

    Wrong name: use the real exported name.

    line 33
  2. Distinctive:

    Create independent new files together, then change existing files one group at a time and run the narrowest check after each group.

    line 44
  3. Distinctive:

    Search for the real symbol before creating a new one.

    line 55
  4. Distinctive:

    If the symbol truly does not exist, decide whether to create it or reuse an existing path.

    line 57

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Inspect Before Build

A design document describes intent. The source describes what exists. Check the source before implementing the plan.

Procedure

1 - Check each plan claim against the source

Claim from planWhat to inspectVerdict
"X has a fake progress bar"X component and its state sourceMay already be real
"Y is minimal"Y componentPlan may be right about missing features
"function N exists"Actual exports and signaturesPlan may name the wrong symbol

Common findings:

  • Already fixed: skip it; do not re-implement.
  • Wrong name: use the real exported name.
  • Already done: the design doc is stale; skip it.
  • Partially right: implement only the true gap.

2 - Record the discrepancies

Record a short "What the plan got wrong" note in the plan or handoff. Future sessions should not rediscover the same stale claims.

3 - Batch independent work, serialize dependent work

Create independent new files together, then change existing files one group at a time and run the narrowest check after each group.

Order within a phase:

  1. New migrations and independent utilities
  2. Existing code that depends on them
  3. Registration, routing, or index wiring

4 - When the build breaks

  • Search for the real symbol before creating a new one.
  • A missing import often means the name changed; find the new name.
  • If the symbol truly does not exist, decide whether to create it or reuse an existing path.

5 - Verify against behavior

A compile or typecheck proves syntax, not the requested behavior. Run the smallest runtime check that exercises the changed path.

Anti-patterns

  • Implementing every line of a plan literally.
  • Trusting a symbol name in a design document without checking exports.
  • Re-implementing a feature that already exists.
  • Treating a successful build as proof the behavior works.
  • Rewriting the plan silently instead of recording the discrepancy.

Verification Checklist

  • Every plan claim about current code was checked against source.
  • Real symbol names and signatures were used.
  • Work already present was skipped.
  • Discrepancies were written back into the plan or handoff.
  • The changed path has a runtime check, not only a compile.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.