sw-task

sw-task is a skill for Claude Code from anton-abyzov/specweave. It costs 46 tokens per session (1,609 once invoked), scanned A, original, MIT.

A task-tracking skill for SpecWeave, a workflow that stores task definitions and state in project files. It uses an append-only JSON Lines ledger so multiple coding agents can share task ownership and progress.

In plain words
What is it for?
Use it to claim tasks, mark them done with evidence, release ownership, skip or block work, and check what task comes next.
Why use it?
It prevents agents from editing the same files at once or losing track of who owns a task. It also keeps a durable record of completed, skipped, blocked, and released tasks.

Skill for Claude Code

Written for Claude Code: argument-hint in frontmatter. Also seen: mentions Codex.

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/anton-abyzov/specweave/sw-task
Any agent
npx skills add anton-abyzov/specweave --skill sw-task
Clone the repo
git clone --depth 1 https://github.com/anton-abyzov/specweave

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for sw-task

README.md
[![agentmods](https://agentmods.dev/badge/skills/anton-abyzov/specweave/sw-task.svg)](https://agentmods.dev/skills/anton-abyzov/specweave/sw-task)
Your own site
<a href="https://agentmods.dev/skills/anton-abyzov/specweave/sw-task"><img src="https://agentmods.dev/badge/skills/anton-abyzov/specweave/sw-task.svg" alt="Measured on agentmods" height="20"></a>
Per session 46 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,609 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00046 $0.01609
Opus 5 $0.00023 $0.00805
Sonnet 5 $0.00009 $0.00322
Haiku 4.5 $0.00005 $0.00161

Measured 2d ago against content hash 80ec0ef20287, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

sw-task scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/sw-task/SKILL.md · 112 lines

How it starts

The opening of the file, as written. The whole thing — 112 lines — stays where its author put it; the contents beside it link to each section on GitHub.

sw-task — the SpecWeave task ledger

Task state lives in .specweave/increments/<id>/ledger.jsonl: an append-only JSONL file that any agent, from any vendor, can write. tasks.md holds the task DEFINITIONS; the ledger holds the STATE. Never edit a ledger line — append.

Two paths, one file format. Run specweave --version: exit 0 → CLI path (accelerator). Command not found → manual path. Both produce byte-identical ledger lines.

Formats (copy verbatim)

Task definition in tasks.md — header + one field line:

### T-01 Fold the ledger
- AC: AC-01, AC-02 | Files: src/a.ts, src/a.test.ts | Test: npm test -- a

Files: is the ownership unit: claim the task that owns a file before editing it.

Ledger line — one JSON object per line, keys in this order:

{"t":"T-01","e":"claim","by":"codex@mbp","at":"2026-09-02T10:00:00Z","note":"optional","evidence":"optional"}
  • t task id · e one of claim done release block skip · by agent id · at ISO-8601 UTC.
  • note is REQUIRED for skip and block. evidence is REQUIRED for done (test result and/or commit sha).
  • UTF-8 without BOM, one object per line, \n-terminated. Blank, malformed and BOM-prefixed lines are skipped and counted, never fatal.
  • Agent id = SPECWEAVE_AGENT if set, else <tool>@<short-hostname> (e.g. codex@mbp).

Derived state (how to read the ledger)

Sort events by at (ties: smaller by first), then apply per task:

Event Accepted when Result
claim task open, or held by you, or the holder's claim is older than the 2h lease claimed by <you>
done has evidence, and you hold it / it is open / the claim is stale done (terminal)
release you are the holder open
block open or held by you blocked + reason
skip has note, not already done/skipped skipped (terminal)

Earliest live claim wins; a claim with no done after 2 hours is stale and may be taken over. A task with no ledger events is open.

Read the full file on GitHub · 112 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 112 lines · 46 tokens per session scan A 80ec0ef20287

Subscribe to this mod's changes

sw-task is a skill published in the GitHub repository anton-abyzov/specweave (159 stars, last pushed today), licensed MIT. It adds 46 tokens to every session and 1,609 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.