workspace-init
Skill2 starsChanged 6 months agoMITView on GitHub
When the user wants to initialize an existing project workspace for Project Skills before project context exists. Use when they have a parent folder with one or more repo folders but no `.agents/`, `ops/`, scratch area, agent instructions, or project context files yet. Scaffolds a clean workspace without moving app code unless explicitly asked.
- Reads credentials
2 other projects use a version of this.
Adopted by
Most recently taken up by
- vines90/foreign-trade-cursor-skillstheir own version, first seen Nov 8, 2026
- patricio0312rev/workspacestheir own version, first seen Jan 25, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- patricio0312rev/workspaces★ 9same name
- vines90/foreign-trade-cursor-skills★ 9same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- decision-captureSkill
- execution-traceSkill
- project-contextSkill
- requirements-from-decisionsSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
19 rules. About 0 are common, 14 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈14
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 17
Use this skill to set up the filesystem shape Project Skills expects when a user is bringing an existing project into an agent-friendly …
- Very specific: line 38
Do not assume .agents/ belongs inside the production repo.
- Distinctive: line 49
Do not use this when:
- Distinctive: line 58
Treat the current directory as the project workspace unless the user names another path.
- Distinctive: line 60
If exactly one repo folder is found, reference it in generated instructions.
- Very specific: line 61
If multiple repo folders are found, list them and write repo references as TBD rather than guessing.
- Distinctive: line 62
Create only safe workspace scaffolding:
- Distinctive: line 70
Do not move, rename, delete, or rewrite the existing repo folder unless explicitly instructed.
- Very specific: line 71
Do not copy raw transcripts, secrets, .env files, or private source dumps into committed paths.
- Very specific: line 72
If Project Skills are already installed under .agents/skills/, preserve them.
- Distinctive: line 73
If skills are not installed, tell the user the exact install command.
- Distinctive: line 77
Before writing files, inspect the workspace:
- Distinctive: line 85
Use shell-equivalent discovery appropriate to the agent environment.
- Distinctive: line 93
If a file already exists, update it conservatively or append a clearly marked Project Skills section.
- Distinctive: line 99
Create this only if needed to keep the folder visible in git.
- Very specific: line 148
If the workspace root is itself a git repo, this .gitignore may be committed.
- Distinctive: line 211
If Project Skills are not already installed, suggest one of these commands.
- Distinctive: line 225
Do not run global install by default.
- Distinctive: line 238
Use this next prompt shape:
≈ “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
Workspace Init
Overview
Use this skill to set up the filesystem shape Project Skills expects when a user is bringing an existing project into an agent-friendly workspace.
This skill runs before project-context. It creates the durable project workspace layer:
<project-workspace>/
.agents/
skills/ # installed Project Skills, if present
source/ # ignored local source drops, optional
project-context.md # created later by project-context
decisions.md # created later by decision-capture
requirements.md # created later by requirements-from-decisions
execution-trace.md # created later by execution-trace
ops/ # curated project/business/source context
scratch/ # local throwaway work
<repo-folder>/ # existing app/repo folder, not moved by default
AGENTS.md # generic/Codex instructions
CLAUDE.md # Claude Code instructions
.gitignore # workspace-level ignore rulesDo not assume .agents/ belongs inside the production repo. The safer default is a parent workspace that contains both .agents/ and the app repo folder.
When to Use
Use this when:
- The user says “init”, “initialize”, “set up project context”, “scaffold Project Skills”, or similar.
- The user has created a parent folder around an existing repo but not the rest of the workspace.
- The user wants to move an existing project into Project Skills without polluting the production repo.
- The user needs
AGENTS.md,CLAUDE.md,.agents/,ops/, andscratch/defaults.
Do not use this when:
.agents/project-context.mdalready exists and the task is to update project context — useproject-context.- The user wants to extract decisions — use
decision-capture. - The user wants requirements or traceability — use the downstream skills.
- The user explicitly wants files created inside the app repo; confirm before doing that.
Default Behavior
- Treat the current directory as the project workspace unless the user names another path.
- Detect existing repo folders by looking for immediate child directories containing
.git. - If exactly one repo folder is found, reference it in generated instructions.
- If multiple repo folders are found, list them and write repo references as
TBDrather than guessing. - Create only safe workspace scaffolding:
.agents/.agents/source/ops/scratch/AGENTS.mdCLAUDE.md- workspace
.gitignore
- Do not move, rename, delete, or rewrite the existing repo folder unless explicitly instructed.
- Do not copy raw transcripts, secrets,
.envfiles, or private source dumps into committed paths. - If Project Skills are already installed under
.agents/skills/, preserve them. - If skills are not installed, tell the user the exact install command.
Workspace Discovery
Before writing files, inspect the workspace:
pwd
find . -maxdepth 2 -name .git -type d
find . -maxdepth 1 -type d | sortUse shell-equivalent discovery appropriate to the agent environment. The goal is to identify:
- workspace root
- existing app/repo folders
- whether
.agents/already exists - whether
AGENTS.mdorCLAUDE.mdalready exists - whether
.gitignorealready exists
If a file already exists, update it conservatively or append a clearly marked Project Skills section. Do not overwrite user instructions.
Files to Create
.agents/.gitkeep
Create this only if needed to keep the folder visible in git. It is optional.
.agents/source/.gitignore
# Local source drops, raw transcripts, exports, and private files.
*
!.gitignoreops/README.md
# Ops Context
Curated project/source context for agents.
Use this folder for summaries, source indexes, scope notes, decision-log exports, and other project context that should sit beside the app repo rather than inside it.
Do not commit secrets, raw transcript dumps, private exports, or files that should remain local-only.scratch/.gitignore
*
!.gitignoreWorkspace .gitignore
Append these entries if missing:
# Project Skills local/source material
.agents/source/
scratch/
# raw exports / transcript dumps
**/*.raw.json
**/*.transcript.txt
**/*.transcript.md
# local environment files
.env
.env.*
*.localIf the workspace root is itself a git repo, this .gitignore may be committed. If the workspace root is not a git repo, still create it as local hygiene.
AGENTS.md
Create or append:
# Agent Instructions
This workspace uses Project Skills.
## Workspace shape
- `.agents/` stores project context artifacts and installed skills.
- `ops/` stores curated project/source context outside the production repo.
- `scratch/` is local throwaway work.
- `<repo-folder>/` is the implementation repo. Verify the actual path before editing code.
## Load order
Before project-context, decision, requirements, execution trace, or delivery closeout work:
1. Read `.agents/project-context.md` if it exists.
2. Read the relevant skill from `.agents/skills/<skill-name>/SKILL.md` if installed.
3. Preserve source references, open questions, and uncertainty.
4. Do not claim implementation, verification, or acceptance without evidence.
## Project Skills
Common workflow:
1. `workspace-init` — scaffold this workspace if needed.
2. `project-context` — create `.agents/project-context.md`.
3. `decision-capture` — create/update `.agents/decisions.md`.
4. `requirements-from-decisions` — create/update `.agents/requirements.md`.
5. `execution-trace` — create/update `.agents/execution-trace.md`.
## Safety rules
- Do not move or rewrite the app repo unless explicitly instructed.
- Do not ingest raw transcripts by default.
- Do not commit secrets, `.env` files, raw exports, or private source dumps.
- Treat tracker status, merged PRs, and local tests as evidence, not acceptance.Replace <repo-folder> with the detected repo folder when exactly one exists. Otherwise leave it as TBD and list candidate repo folders.
CLAUDE.md
Create or append:
# Claude Code Instructions
Read `AGENTS.md` first.
Project Skills live in `.agents/skills/` when installed. If a requested skill is missing, ask the user to install Project Skills or use the checked-out skill source the user points to.
Generated project artifacts should live at the workspace root under `.agents/`, not inside the production app repo unless explicitly requested.Optional Skill Installation
If Project Skills are not already installed, suggest one of these commands.
Install all skills into the current workspace:
npx skills add house-of-giants/projectskills --skill '*'Install all skills for Claude Code and Codex:
npx skills add house-of-giants/projectskills --skill '*' -a claude-code -a codexDo not run global install by default. Project-local install is safer for client/project-specific workflows.
Output Summary
After initialization, summarize:
- workspace path
- detected repo folders
- files/directories created
- files updated instead of overwritten
- whether Project Skills appear installed
- exact next prompt to run
project-context
Use this next prompt shape:
Use the project-context skill to create .agents/project-context.md for this workspace.
Read AGENTS.md first.
Inspect the workspace structure and the implementation repo folder.
Record where source systems, docs, notes, transcripts, tickets, and repos live, but do not ingest full transcripts by default.Common Pitfalls
- Putting
.agents/inside the app repo when the user wanted a parent workspace. - Moving the existing repo folder without explicit instruction.
- Overwriting existing
AGENTS.md,CLAUDE.md, or.gitignorecontent. - Treating raw transcript/source folders as committed context.
- Installing skills globally when project-local install is the safer default.
- Running
project-contextbefore creating a stable workspace shape.
Verification Checklist
Before finishing:
.agents/,ops/, andscratch/exist.- Raw/source material paths are ignored or clearly local-only.
- Existing app/repo folders were not moved or modified unless explicitly requested.
AGENTS.mdandCLAUDE.mdexist or were intentionally skipped.- Workspace
.gitignoreincludes.agents/source/,scratch/, raw transcript/export patterns, and env patterns. - User has a clear next prompt for
project-context.
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.

