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 arch3rPro/dsh-skills --skill minimal-evidence-checksgit clone --depth 1 https://github.com/arch3rPro/dsh-skillsWrote 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/arch3rpro/dsh-skills/minimal-evidence-checks)<a href="https://agentmods.dev/skills/arch3rpro/dsh-skills/minimal-evidence-checks"><img src="https://agentmods.dev/badge/skills/arch3rpro/dsh-skills/minimal-evidence-checks/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/arch3rpro/dsh-skills/minimal-evidence-checks"><img src="https://agentmods.dev/badge/skills/arch3rpro/dsh-skills/minimal-evidence-checks.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 | $0.00076 | $0.00990 |
| Opus 5 | $0.00038 | $0.00495 |
| Sonnet 5 | $0.00015 | $0.00198 |
| Haiku 4.5 | $0.00008 | $0.00099 |
Grade A, and why
minimal-evidence-checks 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 11d 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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Minimal Evidence Checks
A discipline for running the relevant local evidence before a push, without reflexively running the full repository suite. It applies to any repo with checks and a CI system: CI owns exhaustive coverage and the platform matrix; you own the narrow evidence that the outgoing diff would fail.
The defining constraint: match the evidence to the surface. Every behavior change needs the narrowest available check that would fail for its regression — nothing more, and nothing less. Repeating a passing check merely because a commit follows is waste.
Inspect the outgoing change first
Before selecting checks, know the complete scope of what changed:
- Confirm the checkout and branch.
- Determine the base (the merge target, branch parent, or remote head) and inspect the diff against it.
- Account for staged, unstaged, and untracked paths separately.
After a base merge or retarget, re-inspect the combined scope and reassess which checks it can affect.
Select evidence by surface
There is no universal baseline. Choose the narrowest check that would fail for the change's regression:
| Surface | Relevant evidence |
|---|---|
| Behavior in one module | The focused test file or test name for that module |
| A shared contract | That module's tests plus the adjacent modules that consume it |
| Docs, decisions, comments | The documentation/formatting check; full lint when the workflow requires it |
| Model-, UI-, or user-visible output | The focused snapshot or scenario that owns that output |
| Packaging, exports, build config, entrypoints | The build, hygiene checks, and the owning built-artifact smoke |
| External provider or cross-system behavior | The relevant end-to-end target when credentials are available |
Do not manually repeat a passing check just because commit or push follows — in particular, do not re-run a typecheck immediately before pushing if a pre-push hook already runs it.
Focus coverage on the affected source
What ships with it
1 file 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.
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.
- 11d ago First seen · 72 lines · 76 tokens per session scan A 66a923ba7aa8
minimal-evidence-checks is a skill published in the GitHub repository arch3rPro/dsh-skills (4 stars, last pushed 24d ago), licensed MIT. It adds 76 tokens to every session and 990 once invoked, about $0.0004 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 2026-08-31.
Other skills, from other repositories
suede-ai-eval
Suede Labs AI eval design and coverage audit: AI-SPEC, failure-mode rubric with severity scoring, concrete pass/fail eval cases, coverage and infrastructure scores, and mechanical acceptance gates. Use when a change ships LLM, RAG, agent, classifier, prompt, or generated-media behavior, or when asked to write evals…
suede-mcp-qa
Suede Labs AI MCP release QA, scoped to this pack's own server (mcp/suede-skills-mcp.mjs) and its catalog, install, and docs surface. Runs the full JSON-RPC lifecycle against a live server — initialize, notifications/initialized, ping, tools/list, tools/call, resources/list, resources/read, prompts/list, prompts/get …
suede-parity-contract
Suede Labs cross-surface canon discipline: hold one canonical answer across every surface that states it (web, iOS, Android, docs, a second service) by generating a contract from the reference surface and making the others assert against it, so a divergence fails a test instead of reaching a user or an answer engine.…
test-go
Write, review, and improve Go test code for this project. Use whenever generating, reviewing, or modifying Go tests - including when invoked by the Tester agent, the /test prompt, or any test-related request. Covers table-driven tests, subtests, t.Parallel(), test helpers with t.Helper(), error assertions via…
test-ts
Write, review, and run TypeScript/React tests for this Next.js 16 App Router project. Use whenever writing or modifying .test.ts or .test.tsx files, adding test coverage to components, hooks, Server Actions, or utilities, setting up Vitest configuration, or asked about testing strategy. Covers Vitest (the project's…
prove-checks
Prove a passing check was capable of failing before recording it as evidence. Use when a test, CI job, build-and-diff, smoke test or rehearsal comes back green and that green is about to be treated as proof - especially when the check depends on a setup mutation (a sed/awk rewrite, an env var, a secret, a fixture…