Agent skill

Ospec Change

by clawplays in clawplays/ospec

Create or advance a lightweight OSpec change using the classic fast workflow.

MITAuto-check passedAgent Workflows

Install Ospec Change

skills CLI
$ npx skills add clawplays/ospec --skill ospec-change -a claude-code

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

GitHub CLI
$ gh skill install clawplays/ospec ospec-change --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/clawplays/ospec.git skills-src && mkdir -p .claude/skills && cp -r skills-src/assets/global-skills/codex/ospec-change .claude/skills/ospec-change && 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
ospec-change
GitHub stars
452
Token cost
~2.6k tokens
SKILL.md length
1,380 words
Files
3
Skills in repo
3
Repo updated
First seen
Licence
MIT

At a glance

Create or advance a lightweight OSpec change using the classic fast workflow.

  • Works in 6 steps: .skillrc → ospec index query for the relevant… → .ospec/for-ai/change-protocol.md for… → …
  • Agent Workflows work in your project
  • SKILL.md covers Scope, Read Order, Language and Visibility & Decisions, plus 3 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Ospec Change is an agent skill from clawplays/ospec. Create or advance a lightweight OSpec change using the classic fast workflow.

Its SKILL.md is about 2.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 3 other files (for example `agents/openai.yaml` and `skill.yaml`).

It sits in Agent Workflows. It works with Model Context Protocol. The repository describes itself as: Spec-driven, agentic workflow framework for AI coding agents. Turn a request into a verifiable goal loop — plan, act, verify — with durable specs and evidence in your repo. Works… The licence is MIT.

When your agent uses it

  • Agent Workflows work in your project

Example prompts

  • “/ospec-change”

Workflow steps

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

  1. .skillrc
  2. ospec index query for the relevant .ospec/SKILL.index.json entries (root SKILL.index.json in legacy classic projects) — never read the…
  3. .ospec/for-ai/change-protocol.md for nested projects, or legacy for-ai/change-protocol.md
  4. .ospec/changes/active//proposal.md, tasks.md, and state.json for nested projects, or their legacy classic paths
  5. Read verification.md only when entering verification
  6. Read review.md only when entering closeout

What it can do on your machine

Read from SKILL.md and the folder at commit be449f2. 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

    No scripts in the folder and no shell commands in SKILL.md (its code samples are bash).

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

  • Network

    No URLs in SKILL.md.

    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

Ospec Change loads about 2.6k tokens when it runs. Until then it costs about 23 tokens; SKILL.md has 1,380 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~23
When it runs · the whole SKILL.md, loaded when a task matches
~2.6k

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 passed

The automated check found no risky patterns in SKILL.md.

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 clawplays/ospec at commit be449f2, republished under its MIT licence (© clawplays). 1,380 words, ~2,630 tokens.

Download SKILL.mdSave it as .claude/skills/ospec-change/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
ospec-change
description
Create or advance a lightweight OSpec change using the classic fast workflow.
tags
ospec, cli, workflow

OSpec Change

Use this skill for small or routine requirements where the classic OSpec 1.0 change flow is enough.

Scope

This skill is the fast change lifecycle inside an initialized OSpec project: requirement intake, change naming or matching, proposal and task refinement, implementation guidance, verification, archive readiness check, and finalize closeout.

The user owns profile selection. Once the user chooses a change, keep it on the classic flow regardless of complexity, flags, file count, or batch size. Never auto-promote, reject, or replace it with a Goal. Use ospec-goal only when the user explicitly selects a Goal.

Read Order

  1. .skillrc
  2. ospec index query <keyword...> for the relevant .ospec/SKILL.index.json entries (root SKILL.index.json in legacy classic projects) — never read the whole index file, which grows without bound as changes archive
  3. .ospec/for-ai/change-protocol.md for nested projects, or legacy for-ai/change-protocol.md
  4. .ospec/changes/active/<change>/proposal.md, tasks.md, and state.json for nested projects, or their legacy classic paths
  5. Read verification.md only when entering verification
  6. Read review.md only when entering closeout

change-protocol.md is the whole classic contract — the decision-gate ladder, force archive, and every other rule this profile needs are stated there in full. for-ai/execution-protocol.md is the goal controller layer and tells this profile not to open it, so never go there for a change rule. If change-protocol.md itself is missing, use for-ai/ai-guide.md to route back to the protocol for the active profile.

Language

Write every change document and brainstorm you author in the project document language (.skillrc documentLanguage / managed for-ai/ guidance / existing change docs). Never infer that language from product copy, site locale, or an "English-first" requirement, and never mix languages within one change.

Visibility & Decisions

  • Announce-Before-Act: never run the change flow silently. Announce in one line which skill you are using (ospec-change) and the current stage, which ospec command you are about to run and the artifact it writes, and which gate is blocking when progress stops.
  • Brainstorm-First (forked decisions only): raise a gate only for a genuine fork — mutually exclusive API shapes, competing UI approaches, data-model or storage choices, destructive or hard-to-reverse operations, or a scope conflict with what the user asked for. For routine unambiguous changes — an evident-cause bug fix, a mechanical refactor, a docs update, a small addition with one reasonable implementation — do not open a gate or run ospec brainstorm: proceed with the reasonable default and record your assumptions in proposal.md.
  • On a genuine fork, ask one question at a time and never auto-select a recommended option or resolve a gate yourself — recommended is a hint you show the user, never a choice you may take. Present every gate through the capability ladder, in this order: a harness-native question UI when the harness has one (Claude Code AskUserQuestion, Gemini ask_user), otherwise its plan/approval UI (for example Codex plan mode), otherwise the decision report Chat Prompt as plain chat text. You always ask the user and wait for their actual answer; only the presentation differs, and a required pending decision blocks implementation and closeout identically on every harness. Record the gate with ospec execute decision [changes/active/<change>] --id <id> --question "..." --option id:label:impact [--recommended id] [--required] and the answer with --select <option-id> --answered-by user; that shared decision command is the one controller command a change may use. If you did run ospec brainstorm, do not leave it an unanswered template — record each answer with ospec brainstorm resolve [path] --brainstorm <id> --gate <gate-id> --select <option-id> --answered-by user. In Claude Code the managed session hook re-injects this contract at runtime, but it is a convenience for one harness, not the source of the rule: the contract above binds on Codex, Gemini, Grok, OpenCode, Cursor and Copilot too. Full text: for-ai/change-protocol.md.
  • Zero-Setup: the user only describes the change; you run every ospec command yourself and never ask them to type setup or execution commands. In a Claude Code harness, if .claude/settings.json does not yet reference .ospec/hooks/claude/ospec-claude-hook.cjs, run ospec session hook --target claude --apply once (idempotent).
Show full SKILL.md (732 more words)Show less

Required Logic

  1. Inspect repository state first when posture is unclear.
  2. If the repo is not initialized, stop at initialization guidance instead of forcing a change.
  3. If the request is a new requirement, derive a concise kebab-case change name and create it with ospec change <change-name> [path] (ospec new remains a compatibility alias).
  4. If the matching active change already exists, continue it instead of duplicating it.
  5. Keep the work inside the active change container.
  6. Keep proposal.md, tasks.md, state.json, verification.md, and review.md aligned with actual execution.
  7. Do not create design.md, implementation-plan.md, task graphs, worker packets, or Goal review artifacts for changes.
  8. Put batch changes in the queue and execute them sequentially in a shared worktree.
  9. Use OSpec closeout commands instead of inventing a parallel process.
  10. The current AI performs one lightweight review.md review. APPROVED and APPROVED_WITH_CONCERNS may close automatically; PENDING, NEEDS_CHANGES, and BLOCKED stop closeout.
  11. Set the proposal change_type and documentation contract. Bug fixes may record documentation_impact: none with a concrete reason. Features and docs changes require at least one real project, module, API, or user document; a legacy generated archive summary (OSpec no longer produces them) does not count. Update SKILL.md only when module rules or AI usage contracts changed. Index rebuild is automatic. At planning time run ospec docs obligations --apply: the engine derives the located documentation obligations for this change_type and writes each resolved path#section into state.json and the tasks.md checklist, so you never have to search for where to write. A refactor or perf change that genuinely changed no documented behaviour records ospec docs confirm --id <obligation-id> instead of making a cosmetic edit. A fix with no feature document gets an optional obligation and never blocks archiving.
  12. Closeout is automatic when ready: once implementation, verification.md, documentation policy, and review.md are aligned and ospec verify [changes/active/<change>] passes, run ospec finalize [changes/active/<change>] yourself. Do not stop at ospec archive ... --check (it is a preview only) and do not wait for the user to ask before archiving. Closeout uses direct-closeout (archive locally, no PR) and manual merge as defaults — do NOT ask the user about PR, merge, branch, or worktree strategy; uncommitted change/OSpec files in the working tree are normal and do not block archive. Only open a PR if the user explicitly asked. Only pause closeout when a gate genuinely needs a human: a pending required user decision, real blockers reported by verify or archive, or an explicit user request to preview or approve before archiving.
  13. Force archive is never automatic and is never inferred from urgency, a blocker, or a request to "finish". First report the failing gates and every NOT_VERIFIED item to the user, so their acceptance is informed. Only after they explicitly accept that incomplete work, run ospec finalize [changes/active/<change>] --force-archive --confirm-force-archive <exact-change-name> --reason "<accepted risk>"; the CLI enforces the exact-name confirmation and a non-empty reason itself. Preserve the failed and NOT_VERIFIED evidence: the archive stays incomplete and accepted-risk and must never be presented as completed behavior. Full text: for-ai/change-protocol.md.
  14. Activated optional steps must appear in tasks.md and verification.md, and each one that passed must be listed in the verification.md frontmatter field passed_optional_steps — archive validates that field and blocks while an activated step is missing from it.

Commands

bash
ospec status [path]
ospec change <change-name> [path]
ospec changes status [path]
ospec progress [changes/active/<change>]
ospec verify [changes/active/<change>]
ospec archive [changes/active/<change>] --check   # optional preview only — do not stop here
ospec finalize [changes/active/<change>]          # run automatically once verify passes and no human gate is pending
ospec finalize [changes/active/<change>] --force-archive --confirm-force-archive <exact-change-name> --reason "<accepted risk>"   # explicit user exception only, after reporting the failing gates

Guardrails

  • Progress checklists track reality: tick each proposal.md acceptance criterion as it is actually verified — archiving blocks while proposal.md has unchecked items. review.md stays your one lightweight review; record its decision and complete its checklist before closeout. Never tick an item whose work was not done.
  • The worktree is serial: closeout blocks on uncommitted files outside the proposal affects and documentation scopes. Declare affects honestly, and when unattributed dirty files appear (for example another session's edits), commit, stash, or isolate them instead of archiving over them.
  • Do not assume dashboard workflows exist.
  • Do not confuse repository initialization with change execution.
  • Do not enter queue mode unless the user explicitly asks for queue behavior.
  • Do not claim completion until implementation, verification notes, and closeout status are aligned.
  • If real project tests exist, run or recommend them separately from ospec verify.
  • Archived changes are frozen evidence: never edit anything under changes/archived/ (state.json, documents, or artifacts). Archive-time metadata is synced from the authoritative sources by the CLI; if something still looks inconsistent, report it instead of rewriting history — the knowledge index derives from the authoritative documents and self-heals its cache.

© clawplays, 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 2 other files in assets/global-skills/codex/ospec-change of clawplays/ospec.

  • SKILL.md
  • agents/openai.yaml
  • skill.yaml

Open the folder on GitHubat commit be449f2

Compare with similar skills

Ospec Change 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.

Ospec Change compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Ospec Change this skillclawplays/ospec452—~2.6kAutomated safety check: PassMIT
Crush Configurationcharmbracelet/crush29k—~3.7kAutomated safety check: PassCustom licence
Orca Run Replayiflytek/skillhub5.2k4 repos~3kAutomated safety check: PassApache-2.0
Ixix-infrastructure/Ix1.1k—~1.2kAutomated safety check: PassApache-2.0
PicoClaw Agentsipeed/picoclaw30k—~7.2kAutomated safety check: NotesMIT
Chatgpt AppsHaohao-end/openagent7911 repos~4.9kAutomated safety check: PassApache-2.0

Similar skills

  • Crush Configuration

    charmbracelet/crush

    Explains how to configure the Crush coding agent with crushrc or crush.json, covering providers, models, LSPs, MCP servers, hooks, permissions and config precedence.

    29k GitHub stars~3.7k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Orca Run Replay

    iflytek/skillhub

    Answers questions about a past agent run from its recording, using causal graphs and replay, instead of reconstructing events from memory.

    5.2k GitHub starsUsed in 4 repos~3k tokens
    Agent WorkflowsAuto-check passed
  • Ix

    ix-infrastructure/Ix

    Answer structural questions about a codebase — what a symbol is, what calls it, what a change breaks, where the hotspots are — from a persistent code graph via the ix CLI, instead of grepping.

    1.1k GitHub stars~1.2k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • PicoClaw Agent

    sipeed/picoclaw

    Answers questions about running and changing PicoClaw, from onboarding and model selection to MCP server setup, skill loading and scheduled jobs.

    30k GitHub stars~7.2k tokensUpdated yesterday
    Agent WorkflowsAuto-check: notes
  • Chatgpt Apps

    Haohao-end/openagent

    Build, scaffold, refactor, and troubleshoot ChatGPT Apps SDK applications that combine an MCP server and widget UI.

    791 GitHub starsUsed in 1 repo~4.9k tokens
    Agent WorkflowsAuto-check passed
  • Chatgpt App Builder

    alpic-ai/skybridge

    Guide developers through creating and updating ChatGPT plugins.

    2.2k GitHub stars~1k tokensUpdated 3 days ago
    Agent WorkflowsAuto-check passed

More from clawplays/ospec

  • Ospec

    clawplays/ospec

    Document-driven OSpec workflow for initialization, change/goal routing, validation, archiving, and durable project knowledge.

    452 GitHub stars~2.6k tokensUpdated 1 mo ago
    Auto-check passed
  • Ospec Goal

    clawplays/ospec

    Create or advance a full OSpec goal using the current document, task graph, worker, review, and evidence workflow.

    452 GitHub stars~6.4k tokensUpdated 1 mo ago
    Auto-check passed

Questions about Ospec Change

What does Ospec Change do?

Create or advance a lightweight OSpec change using the classic fast workflow. Ospec Change is an agent skill from clawplays/ospec. Create or advance a lightweight OSpec change using the classic fast workflow.

When should I use Ospec Change?

Ospec Change fits situations like: agent Workflows work in your project.

How do I install Ospec Change in Claude Code?

Run `npx skills add clawplays/ospec --skill ospec-change -a claude-code`. Or copy the skill folder (assets/global-skills/codex/ospec-change in clawplays/ospec) into .claude/skills/ospec-change in your project. Claude Code loads it when a task matches its description.

How do I install Ospec Change in Codex?

Run `npx skills add clawplays/ospec --skill ospec-change -a codex`. Or copy the skill folder (assets/global-skills/codex/ospec-change in clawplays/ospec) into .agents/skills/ospec-change in your project. Codex loads it when a task matches its description.

Can I use Ospec Change 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 clawplays/ospec --skill ospec-change -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/ospec-change, .gemini/skills/ospec-change, .github/skills/ospec-change and .opencode/skills/ospec-change in your project.

What does Ospec Change need to run?

SKILL.md names no scripts, command-line tools or credentials: Ospec Change is instructions for the agent only.

Does Ospec Change access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Ospec Change safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Ospec Change use?

Ospec Change 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 Ospec Change use?

About 2.6k tokens (SKILL.md is roughly 11k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Ospec Change?

Skills that share tags, products or a category with Ospec Change: Crush Configuration (charmbracelet/crush, 29k stars), Orca Run Replay (iflytek/skillhub, 5.2k stars), Ix (ix-infrastructure/Ix, 1.1k stars) and PicoClaw Agent (sipeed/picoclaw, 30k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Ospec Change?

clawplays (a GitHub user) maintains it in clawplays/ospec, which has 452 GitHub stars. The repository holds 3 skills in this directory. The repository was last updated on August 26, 2026.

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