agentsclimarketplace

Reconcile claude md

Skill TomerAberbach/claude-config/skills/reconcile-claude-md

🤖 My Claude Code config!

Install
npx -y skills add TomerAberbach/claude-config --skill reconcile-claude-md

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.6 KB, 557 tokens by cl100k_base, as published. Nobody here has run it

Update the project's CLAUDE.md to match its current structure and conventions.

Goals

  • Delete guidance about structure or conventions that no longer exist
  • Update outdated commands, paths, and names
  • Add guidance for conventions Claude can't derive from the code

Current CLAUDE.md

cat CLAUDE.md

$ARGUMENTS

Workflow

  1. If the project has no CLAUDE.md, suggest /init and stop
  2. Break CLAUDE.md into individual claims: commands, paths, tool choices, workflow rules, style rules. One claim per bullet
  3. Verify each claim against the repo:
    • Commands: confirm the script, target, or binary exists (package.json scripts, Makefile targets, lockfiles). Don't run mutating commands
    • Paths and names: confirm the files and directories exist
    • Conventions: sample a few recent source files and check they follow it
  4. Reconcile the project structure section (see "Project structure section")
  5. Explore for undocumented conventions worth recording: tool choices not implied by config files, commands with non-obvious flags, layout rules, gotchas a newcomer would trip on
  6. Edit CLAUDE.md: delete stale claims, fix outdated ones, add missing guidance. Match the file's existing structure and tone, and admit only facts that pass "What belongs"
  7. Report each deletion, fix, and addition with the evidence for it. Flag claims you couldn't verify instead of guessing

Project structure section

CLAUDE.md should have a shallow tree-style code block of the layout:

  • Annotate each entry with a one-line purpose comment. The comments are the value; an unannotated listing is derivable and doesn't belong
  • Keep it shallow: top-level directories and key files. Collapse directories with uniform contents into one glob entry (e.g. **/<name>/) annotated with the pattern its contents follow
  • Add the section if missing. Prune entries that no longer exist, rename moved ones, and add new key entries

What belongs

CLAUDE.md earns its tokens only with facts Claude can't cheaply derive:

  • Keep: the annotated structure tree, tool and command choices among alternatives, conventions not enforced by tooling, invisible constraints (deploy quirks, protected files, ordering requirements)
  • Cut: unannotated file listings, anything a config file already states, lint rules the linter enforces, standard tool usage

These aren't exhaustive. Reason from first principles when none fits cleanly.

Gives 0 of the 12 instructions most memory context skills give in 557 tokens

Counted across 674 of the 847 authors here whose files we hold, read 2026-08-06

  • inform the user when setup is completein 21 of 674, across 6 files
  • confirm the draft with the user before writingin 21 of 674, across 6 files
  • update the agent skills block in place if it existsin 21 of 674, across 6 files
  • present findings to the userin 20 of 674, across 5 files
  • write the three docs files from seed templatesin 20 of 674, across 5 files
  • ask the user about each decision one at a timein 19 of 674, across 4 files
  • edit CLAUDE.md if it existsin 18 of 674, across 3 files
  • explore current repo statein 18 of 674, across 3 files
  • do not overwrite user edits to surrounding sectionsin 18 of 674, across 3 files
  • back up the original file before overwritingin 16 of 674, across 8 files
  • keep the memory index under 200 linesin 15 of 674
  • Provide actionable steps and verificationin 13 of 674, across 2 files

Said here and by no other author read

  • break CLAUDE.md into individual claims
  • reconcile the project structure section
  • explore for undocumented conventions worth recording
  • edit CLAUDE.md to fix outdated or missing guidance
  • report each change with supporting evidence
  • flag unverified claims instead of guessing

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.