agentleFS
bks-labbks-lab/open-bridge/skills/workspace/SKILL.md

workspace

Skill13 starsChanged 13 days agoMITView on GitHub

Binds code repos and config overlays into a named workspace: shared identity in ~/.workspaces/, repo-local clones under .bridge/ with lockfiles. `workspace create` seeds it; `subscribe --role code` clones a repo; `subscribe --role config` delegates to /overlay. Trigger: "/workspace", "create a workspace", "add a repo to my workspace", "workspace status", "list workspaces".

19 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use improve, grill and 8 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • AnasX7Skill

    improve

    Survey any codebase as a senior advisor and produce prioritized, self-contained implementation plans for OTHER models/agents to execute. Strictly read-only on source code — never implements, fixes, or refactors anything itself. Use when asked to audit a codebase, find improvement opportunities (bugs, security, performance, test coverage, tech debt, migrations, DX), suggest features or where to take the project next (roadmap, product direction), or generate handoff plans for another agent to implement.

    Also in 3 projects that use this

  • renatorozasSkill

    grill

    Adversarial code review. Don't let me ship until the changes pass your scrutiny.

    Also in 3 projects that use this

  • GitHubSkill

    prd

    Generate high-quality Product Requirements Documents (PRDs) for software systems and AI-powered features. Includes executive summaries, user stories, technical specifications, and risk analysis.

    Also in 3 projects that use this

  • genomewalkerSkill

    learn

    Learn from web sources, generate SKILL.md, and store key patterns in soul memory. Trigger keywords: learn, skill-gen, web-learn, documentation, tutorial.

    Also in 3 projects that use this

  • OpenAISkill

    audit

    Audit or critique a product flow, journey, workflow, funnel, onboarding path, checkout path, settings path, screen, or multi-step product experience by capturing screenshots first, then reporting UX, design, and accessibility findings inline from that evidence. Use Figma only when the user explicitly asks for a board. Use when the user asks to audit, review, critique, inspect, assess, analyze, evaluate, or give feedback on a product experience.

    Also in 3 projects that use this

  • VercelSkill

    setup

    Set up Vercel CLI and project configuration. Use when the user says "set up Vercel", "configure Vercel", "link to Vercel", or "vercel init".

    Also in 4 projects that use this

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 17 edited versions we read keep these.

  • 0

    Distinctive

    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
  1. Very specific:

    All commands run through scripts/workspace.py.

    line 54

“Common” is measured: kept by at least 6 in 10 of the 17 edited versions of this file we could read. “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

Workspace — bind repos + config into one named project container

A workspace is a named project identity that pulls together the code repos and config overlays you work on as one unit. It has two disjoint halves:

  • Identity (which project is this) — name, member roots, git remotes — lives in the machine-global shared registry ~/.workspaces/workspaces.json, a tool-neutral file any conformant tool can read/write under a documented lock + atomic-replace protocol. This is how a workspace means the same thing across tools on the machine.
  • Materialization (how does THIS repo realise it) — the actual clones, the resolved-SHA lockfile, the fork-safety exclude block — is repo-local and never leaves the repo.

The engine is scripts/workspace.py (standalone, stdlib-only) — a shared repo-root utility shipped with the Bridge repo itself, not a file inside this skill's own directory. This skill is the /workspace surface over it. Run the referenced file ONLY when the decision tree sends you there.

/workspace is the umbrella; for the deep org-overlay subscription flow (manifest validation, per-file plan, 3-way merge, sync/diff) it delegates to and cross-references /overlay (the bridge-overlay skill). Binding a config overlay via workspace subscribe … --role config calls that same engine (overlay.py) under the hood.

When to use

  • "Make a workspace for project X" → create
  • "Add this repo to the workspace" / "bind these repos together" → subscribe … --role code
  • "Pull the org's shared config into this workspace" → subscribe … --role config (delegates to /overlay)
  • "What workspaces do I have / are the members drifted?" → list / status
  • "Drop a repo / overlay from the workspace" → unsubscribe

NOT for:

  • Managing an org overlay subscription on its own (manifest, sync, diff, authoring) → /overlay (bridge-overlay)
  • Pushing YOUR changes upstream → /promote, /bridge-sync
  • A knowledge-base index over a folder → a provider skill, if your instance ships one (unrelated to this "workspace")

Command surface (the real engine CLI)

All commands run through scripts/workspace.py. Mutating verbs (create, subscribe, unsubscribe, archive, restore, delete) require a user/* branch (see Safety).

IntentCommand
Create a workspaceworkspace create <name> [--dir <path>] [--title <label>] [--description <text>]
One workspace with its effective rulesworkspace show <name> [--json]
Set aside, bring backworkspace archive <name>, workspace restore <name>
Remove, a copy stays under .bridge/trash/workspace delete <name> [--dry-run]
Take an area out of the configworkspace area-remove <area>
Bind a code member repoworkspace subscribe <name> <git-url> --role code [--ref <branch>]
Bind a config overlayworkspace subscribe <name> <git-url> --role config [--precedence <N>]
Remove a member (code or overlay)workspace unsubscribe <name> <member>
List all workspacesworkspace list
Validate definition(s)workspace validate [<name>]
Member drift status (read-only)workspace status [<name>]

add-repo / remove-repo are accepted aliases for subscribe / unsubscribe. --role defaults to code. Run any subcommand with -h for its exact flags.

Decision tree:

  • create / subscribe / unsubscribe (a mutating verb) → read references/mechanics.md for the exact per-flow steps, guards, and exit codes before running.
  • understand the model (identity vs materialization, the shared registry, standalone guarantee, how a second tool co-writes safely) → read references/model.md.
  • list / validate / status (read-only) → just run it; no reference needed.

What each mutating verb produces

  • create <name> → writes the source-of-record definition workflow/workspaces/<name>.yaml (schema_version, id, title, overlays: [], repos: []), then additively mirrors the (thin) identity into the shared registry.
  • subscribe … --role code → clones the member into .bridge/workspaces/<name>/<member>/, records it in the definition + workspaces.lock.yaml (resolved SHA), refreshes the .git/info/exclude guard, and re-publishes the grown identity (directories[], git_remotes[]).
  • subscribe … --role config → delegates to overlay.py (subprocess, never imported), then indexes the newly-materialized overlay name(s) in the definition's overlays[] and the identity's extensions["open-bridge"].
  • unsubscribe <name> <member> → code-first dispatch: removes the clone + lock entry (rebuilding the exclude guard, or dropping it when no code members remain), or delegates overlay removal to overlay.py; then re-publishes the shrunken identity.

Safety (never bypass)

  • Branch gate — every mutating verb refuses unless on a user/* branch (require_user_branch), so CORE/definition data never lands on a shared branch. Read-only verbs (list/validate/status) are always allowed.
  • Trust guard — a member git-url is refused before any network access if it uses a scheme outside {https, ssh, file}, a :: remote-helper transport, or a leading - (argv injection). A bare provider name is an inert seam → exit 3, no lookup (this Bridge is standalone).
  • Public-fork safety — foreign code clones live under .bridge/ and are masked by BOTH the tracked .gitignore and an untracked, per-workspace .git/info/exclude block — two independent guards, so a public fork can never git add -A-publish someone else's repo.
  • Warns-never-fails — the shared-registry write is a post-success side-effect; if it fails (registry newer than supported, I/O error, module absent) the local workspace command STILL succeeds. The identity mirror is additive, never the source of record.

Exit codes: 0 ok · 1 guard refusal (branch/trust/not-found/conflict) · 2 usage · 3 provider-name seam (no standalone resolver) · 4 registry version too new (at the registry CLI only; swallowed to a warning here).

Relationship to /overlay and to other tools

  • /overlay (bridge-overlay) owns the org-overlay subscription lifecycle end-to-end. /workspace uses that same engine to bind a config overlay into a named workspace; for standalone overlay management (sync, diff, authoring) use /overlay directly.
  • Any conformant co-writer (e.g. a launcher or file manager that reads ~/.workspaces/ — see the descriptive docs/schemas/workspaces-registry.schema.yaml) shares the identity: it sees this workspace's row, and its own data lives untouched in its own extensions[<tool>] slice. /workspace never imports or requires such a tool — the interop is a file contract, not a dependency.

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.