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 agentmods add skills/h0x91b/dev-3.0/verify-changesnpx skills add h0x91b/dev-3.0 --skill verify-changesgit clone --depth 1 https://github.com/h0x91b/dev-3.0What 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 | $0.00124 | $0.01683 |
| Opus 5 | $0.00062 | $0.00842 |
| Sonnet 5 | $0.00025 | $0.00337 |
| Haiku 4.5 | $0.00012 | $0.00168 |
Grade A, and why
verify-changes 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 — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
verify-changes — testing and verification in dev-3.0
The two hard gates (lint + touched tests before push, full suite before a PR) live in
AGENTS.md and apply whether or not you read this file. Everything here is the detail
behind them: which runner covers what, how to write a test that fits, and what "enough"
means.
Which config runs what
Vitest with happy-dom and React Testing Library. Three configs, three independent
processes:
| Config | Covers | Script |
|---|---|---|
vitest.config.ts |
renderer (src/mainview/) |
part of bun run test |
vitest.config.bun.ts |
backend (src/bun/) |
bun run test:bun |
vitest.config.cli.ts |
CLI (src/cli/) |
bun run test:cli |
bun run test # mainview + bun + cli in parallel, minus 3 slow e2e files (~6s)
bun run test:full # everything incl. slow e2e (~42s) — CI/PR only
bun run test:watch # watch mode
Running vitest directly (outside bun run): bunx vitest run, never npx.
Narrowing while iterating: bunx vitest run <path>, -t "<test name>", --repeat <n> for
suspected flakes. Because the three configs are separate processes, a failure in one says
nothing about the others — read which prefix ([mainview] / [bun] / [cli]) failed.
Local E2E policy: do not run the full E2E suite locally — bun run test:full and any
equivalent unfiltered command are reserved for CI/PR validation. Investigating a specific
behavior means running that one E2E file or test case.
What to test
- Unit (mandatory) — state reducer actions and their edge cases, all pure functions/utils/parsers, every RPC handler (happy path plus 2-3 error cases), CLI commands (parsing, validation, output), data-layer CRUD plus corrupt-data handling, git operations with mocked spawn, i18n interpolation and pluralization for every locale.
- Component (mandatory) — every major interactive component: board views, task cards, modals, settings panels.
- E2E (CLI-based) — full lifecycle through the CLI plus Unix socket against a real app process in a tmpdir: task lifecycle (create → move statuses → complete), project CRUD, worktree creation and cleanup, notes CRUD, CLI context auto-detection, concurrent writes with no data corruption.
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 First seen · 131 lines · 124 tokens per session scan A f3c78bb94e88
verify-changes is a skill published in the GitHub repository h0x91b/dev-3.0 (248 stars, last pushed 3d ago), licensed Apache-2.0. It adds 124 tokens to every session and 1,683 once invoked, about $0.0006 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 skills, from other repositories
test-patterns
Reference for writing tests that actually protect against regressions — deep assertions over spies, table-driven tests, fixture hygiene, AAA structure. Use when writing, reviewing, or refactoring tests.
writing-tests
Write unit tests, component tests, and integration tests for AiderDesk using Vitest and React Testing Library. Use when creating new tests, adding test coverage, configuring mocks, setting up test files, or debugging failing tests.
autoprompt
Explicit-only useful-first orchestration. Invoke only when the user names autoprompt - typed as /autoprompt or in plain language such as "act in autoprompt mode" - to turn a mission into one executable roadmap, build dependency-safe lanes, and verify the result with independent reviewers. Do not infer invocation from…
autoprompt
Explicit-only useful-first orchestration. Invoke /autoprompt to turn a mission into one executable roadmap, build dependency-safe lanes, and verify the result with independent reviewers. Never infer invocation from ordinary requests. Never resume from leftover artifacts without an explicit resume instruction.
autoprompt
Explicit-only useful-first orchestration. Invoke /autoprompt to turn a mission into one executable roadmap, build dependency-safe lanes, and verify the result with independent reviewers. Never infer invocation from ordinary requests. Never resume from leftover artifacts without an explicit resume instruction.
autoprompt
Explicit-only useful-first orchestration. Invoke /autoprompt to turn a mission into one executable roadmap, build dependency-safe lanes, and verify the result with independent reviewers. Never infer invocation from ordinary requests. Never resume from leftover artifacts without an explicit resume instruction.