Agent skill

Spec Write Tasks

by leo-kuang-ai in leo-kuang-ai/spec-first

Public workflow entrypoint (spec-write-tasks): compile a settled local spec-plan into an optional derived task pack for spec-work, or validate an existing local task pack before execution.

MITAuto-check passedAgent Workflows

Install Spec Write Tasks

skills CLI

$ npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a claude-code

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

GitHub CLI

$ gh skill install leo-kuang-ai/spec-first spec-write-tasks --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/leo-kuang-ai/spec-first.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/spec-write-tasks .claude/skills/spec-write-tasks && 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/

Paste into Claude Code

Install the "spec-write-tasks" agent skill from https://github.com/leo-kuang-ai/spec-first/tree/master/skills/spec-write-tasks into .claude/skills/spec-write-tasks/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spec-write-tasks", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/leo-kuang-ai/spec-first/tree/master/skills/spec-write-tasks

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install leo-kuang-ai/spec-first spec-write-tasks --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/leo-kuang-ai/spec-first.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/spec-write-tasks .agents/skills/spec-write-tasks && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "spec-write-tasks" agent skill from https://github.com/leo-kuang-ai/spec-first/tree/master/skills/spec-write-tasks into .agents/skills/spec-write-tasks/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spec-write-tasks", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install leo-kuang-ai/spec-first spec-write-tasks --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/leo-kuang-ai/spec-first.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/spec-write-tasks .cursor/skills/spec-write-tasks && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "spec-write-tasks" agent skill from https://github.com/leo-kuang-ai/spec-first/tree/master/skills/spec-write-tasks into .cursor/skills/spec-write-tasks/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spec-write-tasks", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/leo-kuang-ai/spec-first.git --path skills/spec-write-tasks

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install leo-kuang-ai/spec-first spec-write-tasks --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/leo-kuang-ai/spec-first.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/spec-write-tasks .gemini/skills/spec-write-tasks && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "spec-write-tasks" agent skill from https://github.com/leo-kuang-ai/spec-first/tree/master/skills/spec-write-tasks into .gemini/skills/spec-write-tasks/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spec-write-tasks", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install leo-kuang-ai/spec-first spec-write-tasks

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/leo-kuang-ai/spec-first.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/spec-write-tasks .github/skills/spec-write-tasks && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "spec-write-tasks" agent skill from https://github.com/leo-kuang-ai/spec-first/tree/master/skills/spec-write-tasks into .github/skills/spec-write-tasks/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spec-write-tasks", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install leo-kuang-ai/spec-first spec-write-tasks --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/leo-kuang-ai/spec-first.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/spec-write-tasks .opencode/skills/spec-write-tasks && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "spec-write-tasks" agent skill from https://github.com/leo-kuang-ai/spec-first/tree/master/skills/spec-write-tasks into .opencode/skills/spec-write-tasks/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spec-write-tasks", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: spec-write-tasks
GitHub stars: 107
Token cost: ~2.6k tokens
SKILL.md length: 1,146 words
Files: 32 (incl. references)
Skills in repo: 35
Repo updated: Oct 8, 2026
First seen: Oct 7, 2026
Licence: MIT

At a glance

Public workflow entrypoint (spec-write-tasks): compile a settled local spec-plan into an optional derived task pack for spec-work, or validate an existing local task pack before execution.

Works in 10 steps: spec-plan is always the single source of… → A task pack may rearrange execution… → A task pack is not progress state,… → …
Explicit plan-splitting/task-doc requests
SKILL.md covers Purpose, Workflow Contract Summary, Scenario Capability and Core Rules, plus 7 more sections
Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Spec Write Tasks is an agent skill from leo-kuang-ai/spec-first. Public workflow entrypoint (spec-write-tasks): compile a settled local spec-plan into an optional derived task pack for spec-work, or validate an existing local task pack before execution. Use for explicit plan-splitting/task-doc requests or high-complexity work suitability; do not use for plan authoring, implementation execution, unresolved scope, small low-risk plans, progress/approval state, remote/generic task lists, or generated runtime mirror edits. Keep the plan as single source of truth; tasks are derived…

Its SKILL.md is about 2.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 37 other files, including reference files (for example `agents/openai.yaml`, `evals/README.md` and `evals/boundary-cases.json`).

It sits in Agent Workflows, covering Task breakdown. The repository describes itself as: 仓库原生 AI Coding Harness —— 把一次性 AI 对话变成可治理、可验证、可沉淀的工程闭环 · spec-first.cn. The licence is MIT.

When your agent uses it

Explicit plan-splitting/task-doc requests
High-complexity work suitability
Do not use for plan authoring
Implementation execution

Example prompts

“/spec-write-tasks”

Workflow steps

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

spec-plan is always the single source of truth.
A task pack may rearrange execution slices, but must not change scope, acceptance criteria, non-goals, repo ownership, or product decisions.
A task pack is not progress state, approval state, a lifecycle database, or a second plan.
context_refs are bounded reading pointers, not scope authority. Whole-plan or whole-directory refs are low-quality unless paired with…
If parent-workspace scope is present, the plan or task pack must name target_repo at top level or per task. Missing repo scope returns to…
Source orientation is advisory and bounded: source plan sections, plan-declared files, nearby tests, and precise local contracts only when…
Scripts validate identity, freshness, structure, hashes, concrete paths, and same-wave overlap. LLM/reviewers judge semantic task quality.
Do not hand-edit .claude/, .codex/, or .agents/skills/ as source fixes.
repo selects the artifact/source resolution root only. It does not authorize or select the downstream mutation target_repo.
Compile executable tasks only from status: active. completed, partially-shipped, and superseded plans return source_plan_non_active; do…

What it can do on your machine

Read from SKILL.md and the folder at commit 74655dc. 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.
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

Spec Write Tasks loads about 2.6k tokens when it runs, and up to ~16k if it reads all its reference files. Until then it costs about 137 tokens; SKILL.md has 1,146 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~137
When it runs · the whole SKILL.md, loaded when a task matches: ~2.6k
With references · SKILL.md plus every file in references/, read only if the agent opens them: ~16k

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 leo-kuang-ai/spec-first at commit 74655dc, republished under its MIT licence (© leo-kuang-ai). 1,146 words, ~2,551 tokens.

Download SKILL.mdSave it as .claude/skills/spec-write-tasks/SKILL.md (or your agent's skills folder). This skill also uses 31 other files; get the full folder from GitHub.

name: spec-write-tasks
description: Public workflow entrypoint (spec-write-tasks): compile a settled local spec-plan into an optional derived task pack for spec-work, or validate an existing local task pack before execution. Use for explicit plan-splitting/task-doc requests or high-complexity work suitability; do not use for plan authoring, implementation execution, unresolved scope, small low-risk plans, progress/approval state, remote/generic task lists, or generated runtime mirror edits. Keep the plan as single source of truth; tasks are derived and optional.

`spec-write-tasks`

spec-write-tasks is an optional derived layer between spec-plan and spec-work. It does not execute code. It either compiles a settled source plan into a derived task pack, validates an existing task pack, or returns a no-task-pack decision.

It is a public workflow with the unified entrypoint spec-write-tasks.

Purpose

Decide whether a derived task pack is worthwhile, compile one when it reduces execution risk or context load, and validate existing task packs before they reach spec-work.

Workflow Contract Summary

When To Use

Use when a settled local source plan is too large or dependent for direct execution, when the user explicitly asks to split a plan into tasks, or when an existing local task pack needs validation before spec-work.

When Not To Use

Do not use for unresolved product or architecture scope, small changes that do not need a derived layer, remote repositories or package names, marketplace skill lookup, or implementation prompts that should go directly to spec-work.

Inputs

A local source plan path, existing task-pack path, or clear task-splitting request that resolves to one local source plan. Inputs must stay local and source-owned; remote repos, packages, marketplace identifiers, and generic task lists are rejected.

Outputs

One of: executable derived task pack, validated existing task-pack handoff, skip, return-to-plan, draft-only, or validate-only envelope. Task packs are derived execution indexes and never replace the source plan.

Artifacts

Executable task packs live under docs/tasks/ and must carry an artifact-root-relative POSIX source_plan, canonical source_plan_hash, generated_by: spec-write-tasks, mode: derived, and a valid Task Pack Contract JSON block. Executable identity is source-plan-path+body-hash; spec_id is an optional compatibility trace rather than a required identity field.

Failure Modes

Use machine-readable reason codes: source_plan_missing, ambiguous_plan, wrong_chain, stale_hash, unverifiable_hash, invalid_contract, repo_scope_missing, or scope_gap.

Workflow

Classify input, verify plan identity and repo scope, choose one branch (compile, skip, return-to-plan, draft-only, validate-only), use bounded source orientation only when it improves task boundaries, write or validate the task pack, then return the final decision envelope.

Downstream Consumers

spec-work, high-risk spec-doc-review handoff, human reviewers, and later code-review/compound workflows.

Scenario Capability

Follows docs/contracts/workflows/scenario-capability-matrix.md (default).

Overrides: none

Core Rules

spec-plan is always the single source of truth.
A task pack may rearrange execution slices, but must not change scope, acceptance criteria, non-goals, repo ownership, or product decisions.
A task pack is not progress state, approval state, a lifecycle database, or a second plan.
context_refs are bounded reading pointers, not scope authority. Whole-plan or whole-directory refs are low-quality unless paired with narrower anchors.
If parent-workspace scope is present, the plan or task pack must name target_repo at top level or per task. Missing repo scope returns to spec-plan.
Source orientation is advisory and bounded: source plan sections, plan-declared files, nearby tests, and precise local contracts only when they improve task boundaries.
Scripts validate identity, freshness, structure, hashes, concrete paths, and same-wave overlap. LLM/reviewers judge semantic task quality.
Do not hand-edit .claude/, .codex/, or .agents/skills/ as source fixes.
--repo <artifact-root> selects the artifact/source resolution root only. It does not authorize or select the downstream mutation target_repo.
Compile executable tasks only from status: active. completed, partially-shipped, and superseded plans return source_plan_non_active; do not hash them into a new executable task pack or route them to spec-work. A legacy plan without managed status remains a visible source-plan-lifecycle-unmanaged limitation.

Input Paths

Source plan path: read the source plan frontmatter and focused sections: Requirements, Scope Boundaries, Technical Approach, Implementation Units, Files, Test Scenarios, Verification, and Deferred to Implementation.
Existing task-pack path: read the task pack, source plan, and Task Pack Contract; validate identity/freshness/structure before treating it as executable.
Bare task-splitting request: resolve exactly one local source plan; if resolution is ambiguous, ask for the path.

Use Task Quality Guide for accepted/rejected input classes, task-ready source plan rules, source orientation, traceability, vertical slices, context_refs, done_signal, stop_if, and review-intent quality guidance.

Branches

compile: plan is task-ready and a task pack reduces execution risk or context load.
skip: direct spec-work is cheaper and safer than adding a derived layer.
return-to-plan: scope, acceptance, architecture, verification, or repo ownership is unresolved.
draft-only: a non-executable outline may help discussion but cannot be handed to spec-work.
validate-only: an existing task pack is checked before execution.

Use Execution Handoff Contract for branch outputs, final envelope fields, high-risk review handoff, hash rules, and failure reason codes.

Show full SKILL.md (431 more words)Show less

Task Pack Output

When compiling, read Task Pack Schema and write the task pack body with:

Overview
Source Summary
Traceability Matrix
Task Graph
Execution Waves
Task Pack Contract
Task Cards
Orientation Evidence
Validation Notes
Regeneration Rules

The Task Pack Contract fenced JSON block is the machine-readable source for validators. Human-readable sections compress context for executors and reviewers, but deterministic validation proves only identity, freshness, and structure.

Final Decision Envelope

Every run must end with the compact envelope from Execution Handoff Contract.

Required posture:

include decision, reason_code, identity_basis, artifact_root, source_plan, task_pack, task_pack_validity, deterministic_handoff, validity_scope, semantic_posture, dispatch_authorization, validation, orientation, and next_action;
🔴 GATE (verification): run spec-first tasks validate <task-pack-path> --repo <artifact-root> --json before reporting deterministic_handoff or validation fields;
run spec-first tasks hash <plan-path> --repo <artifact-root> --json when computing or comparing source_plan_hash, and copy its portable source_plan field rather than deriving identity from an absolute machine path;
🔴 GATE (verification): never self-report deterministic_handoff: true without CLI JSON evidence;
🔴 GATE (handoff): allow next_action: spec-work-task-pack only for valid deterministic handoff plus semantic posture generated-this-run or reviewed-existing;
🔴 GATE (verification): for semantic_posture: reviewed-existing, the envelope must carry evidence metadata or a verifiable review-outcome reference; a bare reviewed-existing claim without current evidence is not eligible for next_action: spec-work-task-pack;
🔴 GATE (handoff): for dispatch_authorization: authorized, the envelope must carry a bounded continuation reference or doc-review outcome reference; absent that, report dispatch_authorization: missing.

High-risk packs return next_action: review-task-pack with one reason and the copy-ready current-host invocation spec-doc-review mode:headless mutation:report-only output:json roster:full <task-pack-path>。Consumer 必须把 derived task pack 保持 byte-preserving，并从 task_pack_outcome 读取 deterministic intake、semantic review 与 terminal owner；不要仅凭 Review complete 推导 reviewed-existing。roster:full 仍只保留实际 qualified personas，但不会让 high-risk review 因 standard budget 丢失同时成立的 security、scope 或 adversarial lens。Do not auto-dispatch review unless the explicit bounded continuation conditions in Execution Handoff Contract are met.

Portability Boundary

Runtime references must be packaged skill files (references/, agents/) or optional named integration points distributed with spec-first. spec-plan, spec-work, and spec-doc-review are integration names, not required files to read from this package.

evals/ files are maintainer-only validation fixtures; the official .skill packager excludes root evals/, so runtime users must not need them. Repo-local scripts, validation reports, technical plans, project-role documents, historical snapshots, and .spec-first/audits/** are maintainer evidence, not user-runtime dependencies.

Do Not

Do not execute implementation work.
Do not generate task packs for small low-risk plans by default.
Do not invent missing scope, acceptance criteria, repo targets, or product decisions.
Do not turn review_gate into approval state or validator-owned semantic risk.
Do not make task quality analyzer warnings a hard validator gate.
Do not list repo-local validation reports or historical plans as runtime references.

References

Files

SKILL.md and 31 other files (references) in skills/spec-write-tasks of leo-kuang-ai/spec-first.

SKILL.md
agents/openai.yaml
evals/README.md
evals/boundary-cases.json
evals/eval.yaml
evals/expected-behavior-cases.json
evals/failure-cases.json
evals/output-quality-cases.json
evals/output/cases.jsonl
evals/semantic_config.json
evals/skillup/cases/bare-request-asks-path.yaml
evals/skillup/cases/non-active-rejected.yaml
evals/skillup/cases/remote-input-rejected.yaml
evals/skillup/cases/small-plan-skip.yaml
evals/skillup/fixtures/repos
… and 17 more

Open the folder on GitHubat commit 74655dc

Compare with similar skills

Spec Write Tasks 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.

Spec Write Tasks compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Spec Write Tasks this skillleo-kuang-ai/spec-first	107	—	~2.6k	Automated safety check: Pass	MIT	2 days ago
MemPalace Task HandoffMemPalace/mempalace	59k	—	~1.9k	Automated safety check: Pass	MIT	today
Planning And Task Breakdownabashev/vfs-s3	106	8 repos	~1.9k	Automated safety check: Pass	Apache-2.0	4 days ago
Incremental Implementationaddyosmani/agent-skills	104k	1 repos	~2.3k	Automated safety check: Pass	MIT	7 days ago
ULW Plan Workflowcode-yeongyu/oh-my-openagent	70k	—	~3.9k	Automated safety check: Pass	Custom licence	today
Ask NavigatorYeachan-Heo/oh-my-claudecode	40k	—	~4.1k	Automated safety check: Pass	MIT	2 days ago

Similar skills

MemPalace Task Handoff
MemPalace/mempalace
Creates, hands off, claims, executes and closes agent tasks through the MemPalace logstream, with approval of the exact task before it is recorded.
59k GitHub stars~1.9k tokensUpdated today
Agent WorkflowsAuto-check passed
Planning And Task Breakdown
abashev/vfs-s3
Breaks work into ordered tasks. An agent skill from abashev/vfs-s3.
106 GitHub starsUsed in 8 repos~1.9k tokens
Agent WorkflowsAuto-check passed
Incremental Implementation
addyosmani/agent-skills
Delivers a change in thin vertical slices, each implemented, tested, verified and committed before the next, using vertical, contract-first or risk-first slicing.
104k GitHub starsUsed in 1 repo~2.3k tokens
Agent WorkflowsAuto-check passed
ULW Plan Workflow
code-yeongyu/oh-my-openagent
Explore-first planning that turns a vague or large request into one decision-complete work plan, written only after your approval and executed by a separate worker.
70k GitHub stars~3.9k tokensUpdated today
Agent WorkflowsAuto-check passed
Ask Navigator
Yeachan-Heo/oh-my-claudecode
Charts a foggy effort into a map of decision tickets on the repo's issue tracker and works through them one per session, producing decisions rather than deliverables.
40k GitHub stars~4.1k tokensUpdated 2 days ago
Agent WorkflowsAuto-check passed
Implementation Plan Creator
tailcallhq/forgecode
Writes a structured Markdown implementation plan with checkbox tasks, verification criteria and risks, then checks it with a validation script; no code changes.
7.6k GitHub starsUsed in 1 repo~1.1k tokens
Agent WorkflowsAuto-check passed

More from leo-kuang-ai/spec-first

All 35 skills in this repo

Spec App Consistency Audit
leo-kuang-ai/spec-first
Audit mobile App PRD/Figma/local-source consistency across page routes, KMP/Clean Architecture, components, analytics, i18n, engineering quality, and industry lenses before runtime validation; use…
107 GitHub stars~4.6k tokensUpdated 2 days ago
Auto-check passed
Spec Handoff
leo-kuang-ai/spec-first
Create a durable cross-session handoff or resume from a user-selected continuity source.
107 GitHub stars~1.8k tokensUpdated 2 days ago
Auto-check passed
Spec Pov
leo-kuang-ai/spec-first
Give a decisive, project-grounded verdict on an external input — judged against the current project, not in the abstract.
107 GitHub stars~4.5k tokensUpdated 2 days ago
Auto-check passed
Spec Resolve PR Feedback
leo-kuang-ai/spec-first
Resolve PR review feedback by evaluating validity and fixing issues with conflict-aware resolver dispatch.
107 GitHub stars~1.8k tokensUpdated 2 days ago
Auto-check: notes
Spec Riffrec Feedback Analysis
leo-kuang-ai/spec-first
Analyze explicit Riffrec product-feedback captures, including riffrec-.zip, the Riffrec session.json + events.json + recording.webm + voice.webm bundle, or media/notes the user identifies as a…
107 GitHub stars~1.4k tokensUpdated 2 days ago
Auto-check passed
Spec Compound
leo-kuang-ai/spec-first
Document a recently solved problem or durable project vocabulary in docs/solutions/ or CONCEPTS.md.
107 GitHub stars~18k tokensUpdated 2 days ago
Auto-check passed

Questions about Spec Write Tasks

What does Spec Write Tasks do?

Public workflow entrypoint (spec-write-tasks): compile a settled local spec-plan into an optional derived task pack for spec-work, or validate an existing local task pack before execution. Spec Write Tasks is an agent skill from leo-kuang-ai/spec-first. Public workflow entrypoint (spec-write-tasks): compile a settled local spec-plan into an optional derived task pack for spec-work, or validate an existing local task pack before execution.

When should I use Spec Write Tasks?

Spec Write Tasks fits situations like: explicit plan-splitting/task-doc requests; high-complexity work suitability; do not use for plan authoring; implementation execution.

How do I install Spec Write Tasks in Claude Code?

Run `npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a claude-code`. Or copy the skill folder (skills/spec-write-tasks in leo-kuang-ai/spec-first) into .claude/skills/spec-write-tasks in your project. Claude Code loads it when a task matches its description.

How do I install Spec Write Tasks in Codex?

Run `npx skills add leo-kuang-ai/spec-first --skill spec-write-tasks -a codex`. Or copy the skill folder (skills/spec-write-tasks in leo-kuang-ai/spec-first) into .agents/skills/spec-write-tasks in your project. Codex loads it when a task matches its description.

Can I use Spec Write Tasks 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 leo-kuang-ai/spec-first --skill spec-write-tasks -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/spec-write-tasks, .gemini/skills/spec-write-tasks, .github/skills/spec-write-tasks and .opencode/skills/spec-write-tasks in your project.

What does Spec Write Tasks need to run?

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

Does Spec Write Tasks 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 Spec Write Tasks 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 Spec Write Tasks use?

Spec Write Tasks 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 Spec Write Tasks use?

About 2.6k tokens (SKILL.md is roughly 10k 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 13k tokens, read only when the agent opens those files.

What are the alternatives to Spec Write Tasks?

Skills that share tags, products or a category with Spec Write Tasks: MemPalace Task Handoff (MemPalace/mempalace, 59k stars), Planning And Task Breakdown (abashev/vfs-s3, 106 stars), Incremental Implementation (addyosmani/agent-skills, 104k stars) and ULW Plan Workflow (code-yeongyu/oh-my-openagent, 70k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Spec Write Tasks?

leo-kuang-ai (a GitHub user) maintains it in leo-kuang-ai/spec-first, which has 107 GitHub stars. The repository holds 35 skills in this directory. The repository was last updated on October 8, 2026.

Source: leo-kuang-ai/spec-first on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Install Spec Write Tasks

Facts

At a glance

What it does

When your agent uses it

Example prompts

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

spec-write-tasks

Purpose

Workflow Contract Summary

When To Use

When Not To Use

Inputs

Outputs

Artifacts

Failure Modes

Workflow

Downstream Consumers

Scenario Capability

Core Rules

Input Paths

Branches

Task Pack Output

Final Decision Envelope

Portability Boundary

Do Not

References

Files

Compare with similar skills

Similar skills

More from leo-kuang-ai/spec-first

Related

Questions about Spec Write Tasks

`spec-write-tasks`