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.
npx skills add avatsaev/av-swe-skill --skill av-swegit clone --depth 1 https://github.com/avatsaev/av-swe-skillWrote 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/skills/avatsaev/av-swe-skill/av-swe)<a href="https://agentmods.dev/skills/avatsaev/av-swe-skill/av-swe"><img src="https://agentmods.dev/badge/skills/avatsaev/av-swe-skill/av-swe/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/avatsaev/av-swe-skill/av-swe"><img src="https://agentmods.dev/badge/skills/avatsaev/av-swe-skill/av-swe.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00204 | $0.01776 |
| Opus 5.5 | $0.00082 | $0.00710 |
| Sonnet 5.5 | $0.00041 | $0.00355 |
| Haiku 4.5 | $0.00020 | $0.00178 |
Grade A, and why
av-swe 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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AV-SWE — plan-driven software engineering at scale
Turn a large software engineering effort into a small, ordered, dependency-aware, verifiable task
plan, and execute it reliably. Built by generalizing the clean-room
scope → plan → implement skills for ordinary development where the real source code is present
(no clean-room blinders), plus explicit long-lived project-management operations for driving a big
project over many sessions.
The whole system lives under a plan root (default <project-root>/swe/) and is driven by eight
operations (block/unblock are two directions of one mechanism). Nothing here is a rigid process
for its own sake — the point is that a big project stays plannable, trackable, and consistently
deliverable without the plan rotting.
This file is deliberately a thin index. Each op's full procedure lives in its own file under ops/;
shared conventions (layout, naming, state machine, config schema) live under conventions/. Read the
linked file for the op you're about to run — do not rely on this page alone.
When to use
- The user has a rough feature idea and wants to pin down what it actually does, what it touches, and what goes wrong before anything is specced or planned.
- The user wants to plan a large project, feature set, migration, or refactor.
- The user wants to break a large body of work into sprints and small atomic tasks.
- The user wants to execute a planned chunk of work with confidence it actually builds and verifies.
- The user asks "where are we", "what's next", "is the plan consistent", or wants to move tasks between states.
Operations at a glance
| Op | Effect | Produces / mutates | Full procedure |
|---|---|---|---|
av-swe grill |
Interrogate a rough feature idea into a decided brief (goals, edge cases, integrations) | swe/grills/<name>.md |
ops/grill.md |
av-swe scope |
Capture requirements and current architecture into spec docs | swe/specs/*.md |
ops/scope.md |
av-swe plan |
Derive an ordered sprint + task plan from specs/source | swe/PLAN.md, swe/sprints/sprint-NNN-*/ |
ops/plan.md |
av-swe implement |
Execute one sprint through the state machine, gated on build+test | task moves + *-summary.md |
ops/implement.md |
av-swe status |
Report plan health: open/in_progress/blocked/done, next sprint, gaps | read-only report | ops/status.md |
av-swe validate |
Check plan integrity (ordering, deps, coverage, ids) | read-only report | ops/validate.md |
av-swe block / unblock |
Move a task to/from the blocked state | task file moves | ops/block-unblock.md |
av-swe configure |
Inspect or set project conventions | swe/av-swe.config.json |
ops/configure.md |
What ships with it
16 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- conventions/config.md 2.2 KB
- conventions/layout.md 3.4 KB
- ops/block-unblock.md 514 B
- ops/configure.md 308 B
- ops/grill.md 8.0 KB
- ops/implement.md 2.9 KB
- ops/plan.md 3.1 KB
- ops/scope.md 2.8 KB
- ops/status.md 817 B
- ops/validate.md 878 B
- references/grill-brief-template.md 3.3 KB
- references/plan-template.md 1.7 KB
- references/spec-overview-template.md 1.8 KB
- references/spec-sub-template.md 1.6 KB
- references/summary-template.md 1.2 KB
- references/task-template.md 2.3 KB
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.
- 3d ago Changed · +5 lines · +38 tokens per session fda674ed92de
- 1mo ago Changed 479962f2ce8f
- 1mo ago First seen · 101 lines · 166 tokens per session scan A dfd27dfc0505
av-swe is a skill published in the GitHub repository avatsaev/av-swe-skill (16 stars, last pushed 4d ago), licensed MIT. It adds 204 tokens to every session and 1,776 once invoked, about $0.0008 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
os-big-picture
ALWAYS invoke this skill when the user asks where the project as a whole stands - "where are we", "what's the big picture", "what have we built", "what is in this project", "map the project", "what does this thing even do", "what is stale", "what can we delete", "update the roadmap" - in any language, and whenever a…
ase-task-research
Research a topic in depth for the current or given task plan, by combining Internet/Web search with a grounding in the source artifacts of the project, and append the resulting research report to the "research" attachment of the task, optionally refining the plan by its recommendations. Use when the user wants to…
to-issues
A planning aid that breaks a PRD or technical specification into small implementation issues, each covering a complete, demonstrable piece of work.
relic-expert-handoff
Evidence-first expert knowledge handoff skill. Use when a founder, senior engineer, salesperson, support lead, consultant, or other domain expert is leaving, changing roles, or needs their judgment made reusable. It turns consented documents, chat, meeting notes, interviews, and explicit facts into a traceable handoff…
backlog
EXPERIMENTAL. Use when working the project-wide out-of-scope backlog at .reviews/code/backlog.md and .reviews/security/backlog.md — cleaning it of entries that are already fixed or gone, or picking the next entry and fixing it end-to-end. Not for producing backlog entries — that is the agentic-review skill — and not…
trailhead-view
Skill "trailhead-view" from ToRvaLDz/trailhead, covering load first, in order, routing: verb to engine, render the map: map [map], 🎯 destination and 🗒️ notes.