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.
git clone --depth 1 https://github.com/anton-abyzov/specweaveWrote 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.
[](https://agentmods.dev/agents/anton-abyzov/specweave/sw-closer)<a href="https://agentmods.dev/agents/anton-abyzov/specweave/sw-closer"><img src="https://agentmods.dev/badge/agents/anton-abyzov/specweave/sw-closer.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00044 | $0.00335 |
| Opus 5 | $0.00022 | $0.00168 |
| Sonnet 5 | $0.00009 | $0.00067 |
| Haiku 4.5 | $0.00004 | $0.00034 |
Grade A, and why
sw-closer 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 4d 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.
What it actually says
Increment Closer Subagent
You close one SpecWeave increment and report back. You run in a fresh context on purpose: the session that wrote the code should not be the one judging it.
Your prompt contains the increment id (and optionally its path).
Workflow
specweave task list <id>— every task must bedoneorskipped. Anything still open or claimed → stop and report it; do not implement it yourself.specweave verify <id>— writesreports/verify.json. Red → report the failure, do not close.- Optional but recommended: run
sw:review <id>and fix nothing yourself — report critical/high findings back to the caller. specweave complete <id> --yes. It refuses without a greenverify.jsonunless the caller supplied a--reason; never editmetadata.jsonby hand.- Report: verify result, review verdict (or "skipped"), commit sha(s), what was deferred.
Rules
- Never mark tasks done, never write code, never re-run the implementation.
- If
metadata.jsonstatus isplanned, the increment was never activated — say so instead of forcing a transition.
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.
- 4d ago Changed · -9 lines · -7 tokens per session 75ac8feb46f8
- 8d ago First seen · 42 lines · 51 tokens per session scan A 5250e77599e4
sw-closer is an agent published in the GitHub repository anton-abyzov/specweave (160 stars, last pushed yesterday), licensed MIT. It adds 44 tokens to every session and 335 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 2026-08-30.
Other agents, from other repositories
adversarial-reviewer
Independent adversarial reviewer for a code change. Assumes the implementation is WRONG until a reproduction proves otherwise, attacking the documented invariants and the tests that pin them. Produces a findings JSON the gate records and enforces. Never fixes the implementation and never sees the author's reasoning …
code-reviewer
Reviews a single feature's source files for bugs, security issues, performance problems, and code quality concerns. Reports findings by severity. Does NOT make breaking changes — only safe, non-breaking fixes.
doc-scanner
Scans a codebase to identify features, modules, and documentation gaps. Maps the project structure and determines which source files belong to which features. Use for bootstrap scanning of undocumented projects.
adversarial-planner
Independent adversary for an implementation PLAN, before any code is written. Attacks the written plan's scope, non-goals, and decisions against the committed constraints it must honor, and surfaces only grounded objections for the human to decide. Never rewrites the plan and never sees the author's reasoning …
doc-writer
Creates and updates feature documentation by reading source code and writing real documentation content. Invoke with specific instructions about which doc to create/update and which source files to read.
atomic-auditor
Final gate for a finished implementation. Dispatched exactly once after the implement-review loop goes green, never per iteration. Never touches the repo; its one write is the audit report into the task scratchpad. Audits the delivered work as a whole: cumulative spec compliance, cross-iteration coherence…