Agent skill

Repository Layout and Docs Auditor

by TyrealQ in TyrealQ/q-skills

Audits a repository's file layout and project documentation against a written convention file, then proposes moves, deletions and doc fixes as an approved plan before touching anything.

MITAuto-check: notesAgent Workflows

Install Repository Layout and Docs Auditor

skills CLI
$ npx skills add TyrealQ/q-skills --skill organize -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install TyrealQ/q-skills organize --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/TyrealQ/q-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/organize .claude/skills/organize && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
organize
GitHub stars
108
Token cost
~4.4k tokens
SKILL.md length
2,649 words
Files
2 (incl. references)
Skills in repo
16
Repo updated
First seen
Licence
MIT

At a glance

Audits a repository's file layout and project documentation against a written convention file, then proposes moves, deletions and doc fixes as an approved plan before touching anything.

  • Works in 7 steps: Context: what started the pass, what the… → Findings: grouped under the detector… → Operations: numbered and grouped for… → …
  • Cleaning up an inconsistent or sprawling repository layout
  • SKILL.md covers References, Core Principles, Workflow and Detectors, plus 1 more section
  • Calls git

What it does

This skill runs nine detectors against a repository's naming, layout, documentation model and superseded-content rules, writing every finding and every proposed operation into a plan file before changing a single file. The user approves or rejects each group of operations in that plan; the only exception is a group labeled as fixing a rule already written in the project's own docs, which is applied without asking, while new files and consolidations of duplicated content always wait for approval.

Three sources decide different questions: disk and git ls-files decide what exists, the project's own documentation decides what something is for and which project rules hold, and a bundled conventions.md reference decides naming and layout wherever the project's docs are silent. Any mismatch between these sources is listed in the plan for the user to resolve, except that a conflict between a project rule and conventions.md is resolved in the project's favor and simply noted. Every move and deletion appears as its own listed operation and none applies before approval, and scope is limited to Markdown files whose job is describing the repository, not deliverable prose.

Moves on cloud-synced paths such as Dropbox, OneDrive or iCloud are done by copying with retry, confirming the copy, then removing the source, rather than renaming a folder in place, and git is left to detect the rename afterward rather than using git mv; a case-only rename goes through the same two-step method using a temporary name. The skill ends by handing off to a separate commit or ship skill.

When your agent uses it

  • Cleaning up an inconsistent or sprawling repository layout
  • Standardizing file and folder naming against a written convention
  • Finding and archiving documentation that is stale or superseded
  • Preparing a repository cleanup plan for approval before a commit

Example prompts

  • “Audit this repo's layout and docs against our conventions and show me the plan.”
  • “Find stale or superseded documentation and propose what to archive.”
  • “Check for naming mismatches between our docs and the actual folder structure.”
  • “Move these files safely, this repo lives in a Dropbox-synced folder.”

Requirements

  • git, to detect file moves after they happen

Workflow steps

7 steps, taken from the first numbered list in SKILL.md.

  1. Context: what started the pass, what the repository holds, and the path scope, if any.
  2. Findings: grouped under the detector letters A to I, each with its path and the line or name at issue. A note is a finding listed for the…
  3. Operations: numbered and grouped for approval. Each group lists its moves, deletions, git rm --cached calls, and documentation edits. A…
  4. Out of scope: each finding that will not be acted on, with the reason.
  5. Open questions: each decision the audit cannot settle, with what the answer changes.
  6. Verification: the checks from Step 7 that apply to this plan.
  7. Rollback: how to reverse each group. A deleted tracked file or an edited document comes back with git checkout HEAD -- before a commit; a…

What it can do on your machine

Read from SKILL.md and the folder at commit d8aaee7. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Repository Layout and Docs Auditor loads about 4.4k tokens when it runs, and up to ~7.5k if it reads all its reference files. Until then it costs about 70 tokens; SKILL.md has 2,649 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~70
When it runs · the whole SKILL.md, loaded when a task matches
~4.4k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~7.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NoteMentions a .env fileSKILL.md:101
    one machine or created at run time (an `.env` file, a virtual environment).

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from TyrealQ/q-skills at commit d8aaee7, republished under its MIT licence (© TyrealQ). 2,649 words, ~4,384 tokens.

Download SKILL.mdSave it as .claude/skills/organize/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
organize
description
Audit project structure and documentation, align names and layout to conventions, remove or archive superseded content, and refine READMEs and CLAUDE.md. Use for cleaning up folders, standardizing layout, fixing stale docs, or streamlining documentation before a commit.

Organize Skill

Audits a repository's layout and its project documentation against one set of conventions, using nine detectors, a plan file, and approval by group, and then hands off to /commit or /ship.

References

  • references/conventions.md: the target state the detectors check against. It covers naming, layout, reusable versus project code, the documentation model, the current-state rule, superseded content, and the .gitignore categories.

Core Principles

  • Plan before acting. Write every finding and every proposed operation to the plan file. Change no file until the user has approved the group that contains it; the one exception is a group labeled "fixes a written rule" (see Plan file), which is applied without asking.
  • Take each answer from its source. Disk and git ls-files decide what exists; the project's docs decide what a thing is for and which project rules hold; conventions.md decides what a thing is called and how it is laid out wherever the project's docs state no rule. List every mismatch between these sources in the plan; the user decides which side changes, except in a group labeled "fixes a written rule". Where a rule written in the project's docs conflicts with conventions.md, follow the project rule and list the conflict as a note.
  • Ask for approval per group. The user approves or rejects each group of operations in the plan. A clear violation of a rule already written in the project's own docs is fixed without asking. New files and consolidations of duplicated content always wait for approval.
  • List every move and deletion for approval. Every move and every deletion appears in the plan as its own operation, and none is applied before the user approves it.
  • Edit project documentation only. Among Markdown files, edit only those whose job is describing the repository. Deliverable prose is outside this skill; see Scope.
  • Move files safely on synced paths. Copy the file or folder with retry, confirm the copy, then remove the source. Never rename a folder in place on a cloud-synced path such as Dropbox, OneDrive, or iCloud. Let git detect renames after the move; do not use git mv. A rename that changes only case uses the same method twice, through the temporary name that detector A gives.

Workflow

StepActionReference
1. AuditResolve the repository root with git rev-parse --show-toplevel, and limit the pass to a path argument when one is given. Record git status --short as the baseline for Step 7. List git ls-files and the gitignored paths (git ls-files --others --ignored --exclude-standard --directory), and walk the disk with those paths included so that detector B can see untracked generations. Read the root CLAUDE.md first, then every README.md in the tree and every owner file the docs point to. When the project's docs forbid sending a file's contents to this model (for example proprietary text), use only its name and path, and list any check that needs the contents (a diff, a count) under Open questions with the reason.conventions.md § Documentation model
2. Layout detectorsRun detectors A to D.Detectors A to D; conventions.md § Naming, § Layout, § Superseded content, § .gitignore categories
3. Documentation detectorsRun detectors E to I.Detectors E to I; conventions.md § Documentation model, § Current-state rule, § Layout, § Naming
4. Plan fileWrite the plan to ~/.claude/plans/organize-<repo>-<YYYYMMDD>.md, or to the path the user names, with the seven sections below.Plan file, below
5. ClarifyAsk with AskUserQuestion for each open question, such as the date of a name that lacks part of its date, the destination of an orphan file, or the owner of duplicated content.Step 1; Detectors A, C, and G
6. ApplyApply the approved groups and the groups labeled "fixes a written rule", one group at a time in plan order, and stop at the first error. Within a group, make the moves and deletions before the documentation edits.Core Principles; Plan file, below
7. VerifyRun git status --short and walk the final tree. Confirm that every folder the root map names exists; every folder that conventions.md § Documentation model requires a README for has one, and the root map points to each top-level README; no approved finding is left open; the paths untracked under detector D are gone from git ls-files; and git status --short shows only the baseline and the changes the plan lists. Report any check that fails and leave it open.Plan file, below (Verification)
8. Hand offSummarize what moved, what was deleted, what was untracked, which documentation files changed, and which findings were left out of scope. Name /commit for a single commit, or /ship to also update CHANGELOG.md and push.Scope

With the argument audit, or when the user asks for an audit only, stop after Step 4 and leave the open questions in the plan file unasked.

Plan file

The plan file has seven sections, in this order:

  1. Context: what started the pass, what the repository holds, and the path scope, if any.
  2. Findings: grouped under the detector letters A to I, each with its path and the line or name at issue. A note is a finding listed for the user with no operation proposed. Notes stay in Findings and are not repeated under Out of scope.
  3. Operations: numbered and grouped for approval. Each group lists its moves, deletions, git rm --cached calls, and documentation edits. A documentation edit shows the old and the new text. A new file, such as a missing README, is drafted in full. Each move or rename carries, in the same group, the edit to every project document that names the old path, and to any note in the project's memory folder (~/.claude/projects/<project>/memory/*.md) that names it.
  4. Out of scope: each finding that will not be acted on, with the reason.
  5. Open questions: each decision the audit cannot settle, with what the answer changes.
  6. Verification: the checks from Step 7 that apply to this plan.
  7. Rollback: how to reverse each group. A deleted tracked file or an edited document comes back with git checkout HEAD -- <path> before a commit; a path untracked with git rm --cached is tracked again with git add <path>; an archived folder moves back from _archive/.

An operations group reads like this:

### Group 2: superseded generations (detector B), needs approval

2.1  delete   report_v1.md                         tracked; report.md is current
2.2  move     outputs/models/run_2026-04-14/  ->  outputs/models/_archive/run_2026-04-14/
2.3  edit     outputs/README.md, line 12
     old: "Results are in run_2026-04-14/ and run_2026-05-02/."
     new: "Results are in run_2026-05-02/."

Label each group "needs approval" or "fixes a written rule". Use the second label only for a group of documentation edits with no move, deletion, new file, or consolidation, and that removes no information: no sentence, clause, finding, or recorded reason that the edit does not restate as a current fact; give the file and line where the project states the rule.

Detectors

Run every detector over the whole tree, gitignored paths included, and record each finding in the plan file under its letter. The target state that each detector checks against is in references/conventions.md.

A. Name drift
  • Looks for: folder and file names that break conventions.md § Naming, and names that differ in case or spelling between the docs and disk.
  • Verify: compare git ls-files and a walk of the disk against every name the root map and the READMEs give.
  • Resolve: propose the rename and the matching docs edit as one operation. On a case-insensitive filesystem, rename in two steps through a temporary name (Data/ to data_tmp/ to data/) so that git and the sync client both record the change. When a name lacks part of its date, list it under Open questions instead of proposing a name.
  • Never flag: names that conventions.md § Naming writes in capitals (files a tool reads by a fixed name, owner files, and acronyms the project's docs write in capitals), or a name that follows a naming rule written in the project's docs (list the conflict with conventions.md as a note).
B. Superseded generations
  • Looks for: folders named v1/, v2/, old/, backup/, bak/, deprecated/, or legacy/; siblings named <name>_old, <name>_backup, <name>_v1, or <name>_v2; dated siblings of current content; and a plan or design document whose work is complete, that describes no part of the current project, and that no current doc points to. Search everywhere outside _archive/, including inside gitignored paths such as outputs/.
  • Verify: for dated siblings, apply the test in conventions.md § Superseded content: a sibling is superseded only when the docs and scripts read the later one alone. Diff each remaining candidate against the current folder beside it; when it holds the current content under a wrong name, propose a rename, not a removal.
  • Resolve: delete the old version or move it to _archive/, following the tracking split in conventions.md § Superseded content.
  • Never flag: anything already inside _archive/, or dated siblings that form a collection under conventions.md § Superseded content. List received material that the docs mark as superseded as a note.
C. Orphan files
  • Looks for: files outside gitignored paths whose name and path appear in no root map, README, script, or tracked Markdown file, and which do not sit inside a folder that the root map or a README describes; and files at the root that conventions.md § Layout does not allow, listed as a note when gitignored.
  • Verify: search for the basename and the relative path across git ls-files '*.md' and the scripts (*.py, *.R, *.sh, *.js, *.ts). Then check the file's own folder; the file counts as covered when the root map or a README describes that folder. A covered file that nothing names, in a folder whose README has a file table that omits it, is a note.
  • Resolve: propose a destination among the existing folders based on file type, or ask the user what the file is for. Never delete an orphan without approval.
  • Never flag: files a tool reads by a fixed name and owner files, as conventions.md § Naming defines them, or deliverables.
Show full SKILL.md (1,039 more words)Show less
D. Tracked per-machine state
  • Looks for: tracked paths that match a pattern in conventions.md § .gitignore categories, tracked per-user files inside a root .<tool>/ folder as that section defines them, and a .gitignore missing one of that section's categories.
  • Verify: confirm each path appears in git ls-files, and check whether .gitignore already has a line covering it.
  • Resolve: propose git rm --cached <path>, plus the matching .gitignore line when none exists. The file stays on disk; only the index entry is removed.
  • Never flag: .env.example or other templates meant to be shared.
E. Stale facts
  • Looks for: every path, filename, folder, command, and count named in project documentation.
  • Verify: check each path and name against disk and git ls-files, and each count against the file or script that produces it. For a command, confirm the script exists and read its argument parser to confirm it accepts the flags named.
  • Resolve: correct the doc to match disk. Never change a number without its source; when the source cannot be found, list the count under out of scope.
  • Never flag: placeholder patterns such as <name>_old or YYYY-MM-DD_slug, external paths that the root map names as outside the repository, or a gitignored path that the docs describe as local to one machine or created at run time (an .env file, a virtual environment).
F. History clauses
  • Looks for: a date on a change, and the phrases listed in conventions.md § Current-state rule.
  • Verify: for every date and every phrase found, read the whole sentence and flag it only when it contrasts the current state with an earlier one. The same words used for present behavior ("runs until the queue is empty") describe a current fact.
  • Resolve: rewrite the sentence to state the current fact, as in the example in conventions.md § Current-state rule, or propose deleting the clause when the sentence holds no current fact.
  • Never flag: the exceptions listed in conventions.md § Current-state rule, or the date in a YYYY-MM-DD_slug filename.
G. Duplication
  • Looks for: the same rule or fact stated in two or more files.
  • Verify: compare the statements and confirm they give the same content rather than two related rules.
  • Resolve: choose the owner: the file the project's docs name as the owner of the topic, or, when they name none, the file whose type in the table in conventions.md § Documentation model matches the topic; ask the user when more than one type fits. Keep the statement in the owner and replace every other statement with a pointer in the form that conventions.md § Documentation model gives.
  • Never flag: a pointer that names its owner, or the one-line entry the root map gives a folder whose README holds the full account.
H. Structure
  • Looks for: a document missing the fixed parts of its type; a folder that conventions.md § Documentation model requires a README for, without one; a top-level folder README the root map does not point to, or a unit README the index does not link; an index written as prose; detail in the root map that belongs in a folder README; a unit README that reports a result without the date it was obtained; sibling unit folders with different inside layouts; a root folder whose subfolders do not mirror the project's existing split, or that holds different kinds of file (scripts, documents, media) directly with no subfolders; a section with fewer than two substantive sentences; a wrapper folder around the role folders; a folder holding one item (unit folders and the role folders in conventions.md § Layout excepted); a collection nested by meaning; outputs/ subfolders that do not match the stages under scripts/.
  • Verify: name each document's type from the table in conventions.md § Documentation model and check its fixed parts in order; compare folders against conventions.md § Layout and collections against § Naming.
  • Resolve: propose the missing README, table, or date, drafted in full in the plan file; move root-map detail to the folder README that owns it and leave a pointer; propose the moves that give each folder the layout in § Layout, except that matching outputs/ to scripts/ goes under Out of scope when it would split or merge pipeline folders.
  • Never flag: folders below a gitignored top-level folder or inside _archive/, item folders in a collection keyed by identifier, which the collection's own README covers, or an index README, a table, a code block, or a run line under the two-sentence rule.
I. Prose
  • Looks for: sentences that announce what follows, restatement of a point already made, padding, and figures of speech.
  • Verify: check each candidate against the user's global writing rules (for example, files under ~/.claude/rules/ and ~/CLAUDE.md) when they exist. The user's writing rules take precedence over the categories above; apply each rule only to the kinds of text it names.
  • Resolve: propose the rewrite, with the old and new sentence side by side in the plan file.
  • Never flag: quoted text, code, command lines, or a banned phrase quoted to state a rule.
Scope of E to I

Detectors E, F, G, and I, and the document checks in H, read only Markdown whose job is describing the repository: the root map, folder READMEs, index and unit READMEs, owner files, and AGENTS.md. Anything the root map lists as a deliverable is outside their scope, and so are correspondence and reports written for readers. Detectors A to D still check the names and locations of those files.

Scope

Include

  • Layout and naming alignment between disk, docs, and conventions.md
  • Removal of tracked superseded content and archiving of untracked superseded content
  • Placement of orphan files
  • Untracking per-machine state and extending .gitignore
  • Refining and restructuring project documentation: creating missing READMEs, consolidating duplicated content into one owner, and moving detail from the root map into folder READMEs
  • Updating path references in the project's memory notes (~/.claude/projects/<project>/memory/), listed in the plan with the move they follow

Exclude

  • Deliverable prose: anything the root map lists as a deliverable, correspondence, and reports written for readers
  • Splitting or merging pipeline, data, or script folders without an explicit request
  • Running workbook, build, or pipeline scripts
  • Anything inside .git/
  • Changes inside gitignored paths other than moves into _archive/; a finding there from detectors A to D that needs another change, such as a rename, goes under Out of scope
  • Committing or pushing

© TyrealQ, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (references) in skills/organize of TyrealQ/q-skills.

  • SKILL.md
  • references/conventions.md

Open the folder on GitHubat commit d8aaee7

Compare with similar skills

Repository Layout and Docs Auditor next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Repository Layout and Docs Auditor compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Repository Layout and Docs Auditor this skillTyrealQ/q-skills108—~4.4kAutomated safety check: NotesMIT
Neat-Freak Knowledge CloseoutKKKKhazix/khazix-skills21k—~1.9kAutomated safety check: PassMIT
Simplify Docstobihagemann/turbo409—~3.1kAutomated safety check: PassMIT
Doc-Code Sync Checkfancyboi999/open-tag203—~1.7kAutomated safety check: PassApache-2.0
Sync Docsayutaz/piper-plus230—~1.4kAutomated safety check: PassMIT
Doc SyncJetBrains/ideavim10k2 repos~2.6kAutomated safety check: PassMIT

Similar skills

  • Neat-Freak Knowledge Closeout

    KKKKhazix/khazix-skills

    Brings project docs, agent rule files, authorized memory and leftover workspace files back in line with what the code and runtime actually do at the end of a work session.

    21k GitHub stars~1.9k tokensUpdated 8 days ago
    Agent WorkflowsAuto-check passed
  • Simplify Docs

    tobihagemann/turbo

    Run a multi-agent review of code comments and markdown documentation for unnecessary content, then fix the issues.

    409 GitHub stars~3.1k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Doc-Code Sync Check

    fancyboi999/open-tag

    Reconciles documentation with code at the end of a change or as a periodic audit, following a repo rule that code changes and doc changes land in one commit.

    203 GitHub stars~1.7k tokensUpdated 1 mo ago
    DevelopmentAuto-check passed
  • Sync Docs

    ayutaz/piper-plus

    コミット前にエージェントチームで全ドキュメント (CLAUDE.md / README / CHANGELOG / docs/) を監査し、コード変更に応じて自動更新します。大規模変更時の documentation drift を予防。

    230 GitHub stars~1.4k tokensUpdated yesterday
    DevelopmentAuto-check passed
  • Doc Sync

    JetBrains/ideavim

    Official

    Keeps IdeaVim documentation in sync with code changes. An agent skill from JetBrains/ideavim.

    10k GitHub starsUsed in 2 repos~2.6k tokens
    DevelopmentAuto-check passed
  • Mermaid Diagrams

    jjmartres/opencode

    Helps an agent pick the right Mermaid diagram type and write the syntax for class, sequence, flow, ER, C4, state and other software diagrams.

    133 GitHub starsUsed in 6 repos~1.9k tokens
    DevelopmentAuto-check passed

More from TyrealQ/q-skills

All 16 skills in this repo
  • Q-Infographics

    TyrealQ/q-skills

    Converts a report or other document into a business story and then an infographic image, pausing for your review after each step.

    108 GitHub stars~814 tokensUpdated 16 days ago
    Auto-check: notes
  • Runs exploratory data analysis on tabular data after you confirm each column's measurement level, then writes CSV tables and a narrative summary.

    108 GitHub stars~1.1k tokensUpdated 16 days ago
    Auto-check passed
  • Extracts pixel, video-frame, speech, music and visual-semantic features from image, video and audio files for research datasets, using local tools or the Gemini API.

    108 GitHub stars~2k tokensUpdated 16 days ago
    Auto-check: notes
  • Consolidates BERTopic, LDA or NMF topic output into a theory-driven classification framework and writes the final labels back to an Excel file.

    108 GitHub stars~1k tokensUpdated 16 days ago
    Auto-check passed
  • Generates branded slide deck images from written content, with a content analysis step, a layout catalog and scripts that merge the slides into PowerPoint or PDF.

    108 GitHub stars~1.1k tokensUpdated 16 days ago
    Auto-check passed
  • Commit

    TyrealQ/q-skills

    Stage and commit uncommitted changes with conventional commit messages.

    108 GitHub stars~1.1k tokensUpdated 16 days ago
    Auto-check: notes

Works with

Questions about Repository Layout and Docs Auditor

What does Repository Layout and Docs Auditor do?

Audits a repository's file layout and project documentation against a written convention file, then proposes moves, deletions and doc fixes as an approved plan before touching anything. This skill runs nine detectors against a repository's naming, layout, documentation model and superseded-content rules, writing every finding and every proposed operation into a plan file before changing a single file. The user approves or rejects each group of operations in that plan; the only exception is a group labeled as fixing a rule already written in the project's own docs, which is applied without asking, while new files and consolidations of duplicated content always wait for approval.

When should I use Repository Layout and Docs Auditor?

Repository Layout and Docs Auditor fits situations like: cleaning up an inconsistent or sprawling repository layout; standardizing file and folder naming against a written convention; finding and archiving documentation that is stale or superseded; preparing a repository cleanup plan for approval before a commit.

How do I install Repository Layout and Docs Auditor in Claude Code?

Run `npx skills add TyrealQ/q-skills --skill organize -a claude-code`. Or copy the skill folder (skills/organize in TyrealQ/q-skills) into .claude/skills/organize in your project. Claude Code loads it when a task matches its description.

How do I install Repository Layout and Docs Auditor in Codex?

Run `npx skills add TyrealQ/q-skills --skill organize -a codex`. Or copy the skill folder (skills/organize in TyrealQ/q-skills) into .agents/skills/organize in your project. Codex loads it when a task matches its description.

Can I use Repository Layout and Docs Auditor in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add TyrealQ/q-skills --skill organize -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/organize, .gemini/skills/organize, .github/skills/organize and .opencode/skills/organize in your project.

What does Repository Layout and Docs Auditor need to run?

Going by SKILL.md and its folder, Repository Layout and Docs Auditor needs the command-line tools its instructions call (git). Our summary lists: git, to detect file moves after they happen.

Does Repository Layout and Docs Auditor access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Repository Layout and Docs Auditor safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Repository Layout and Docs Auditor use?

Repository Layout and Docs Auditor is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Repository Layout and Docs Auditor use?

About 4.4k tokens (SKILL.md is roughly 18k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 3.1k tokens, read only when the agent opens those files.

What are the alternatives to Repository Layout and Docs Auditor?

Skills that share tags, products or a category with Repository Layout and Docs Auditor: Neat-Freak Knowledge Closeout (KKKKhazix/khazix-skills, 21k stars), Simplify Docs (tobihagemann/turbo, 409 stars), Doc-Code Sync Check (fancyboi999/open-tag, 203 stars) and Sync Docs (ayutaz/piper-plus, 230 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Repository Layout and Docs Auditor?

TyrealQ (a GitHub user) maintains it in TyrealQ/q-skills, which has 108 GitHub stars. The repository holds 16 skills in this directory. The repository was last updated on September 23, 2026.

Source: TyrealQ/q-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.