coverage-differential SKILL.md skillA

coverage-differential from sshadows/lethal is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add sshadows/lethal/coverage-differential@git:20260827.b14163c

Pins git:20260827.b14163c in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
sshadows/lethal
Path
.claude/skills/coverage-differential/SKILL.md
Versions recorded
3
Latest version
git:20260827.b14163c, committed Aug 27, 2026
First version recorded
Jul 28, 2026
Size
7,043 bytes
Repository stars
10
License
MIT

Run the two-mode differential gate for any change to coverage collection, selection or attribution — one project twice, identical except the coverage mode, compared per mutant on a verified join. Use before claiming such a change is safe, because the frozen live gates cannot detect these regressions.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# The coverage differential gate

## Why the ordinary gates cannot do this

All four frozen gates have GREEN baselines. A change to how coverage is COLLECTED or ATTRIBUTED is
a no-op for them: every test passes either way, so every mutant is scored either way, and the
verdict tables match whatever the attribution did. R55's first candidate fix looked verified for
exactly this reason. So did R58's client half, whose unit suite was fully green while the real call
was hanging for 300 s against a live container.

A regression here does not fail anything. It changes WHICH tests are believed to cover WHICH
mutants, and the only visible consequence is mutants quietly moving to `no-coverage` — which reads
as "the suite does not cover this code", the most reassuring possible misreading.

## Run it

```bash
# fixtures: the probe drives runSession directly and dumps per-test coverage sets
bun scripts/probe-r58-differential.ts --project fixtures/sandbox-app \
  --tests fixtures/sandbox-tests --mode procedure --out <scratch>/a.json
bun scripts/probe-r58-differential.ts --project fixtures/sandbox-app \
  --tests fixtures/sandbox-tests --mode fenced --out <scratch>/b.json
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add sshadows/lethal/coverage-differential@sha256:7a53074abdcc3c85

The line at the top of this page pins the label git:20260827.b14163c, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_wnimg4g247wc2rwy.svg)](https://markdownregistry.com/a/art_wnimg4g247wc2rwy)

0 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 3 versions of this file, between Jul 28, 2026 and Aug 27, 2026, about one every 15 days.

versioncommittedcommitsizeaudit
git:20260827.b14163c latest2026-08-27 b14163c 7,043 BA view · diff
git:20260820.185caa72026-08-20 185caa7 5,397 BA view · diff
git:20260728.06559562026-07-28 0655956 5,301 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (7043 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

sshadows/lethal · 10 stars · license MIT · pushed 2026-11-07 · branch master

API

GET https://markdownregistry.com/api/v1/artifacts/art_wnimg4g247wc2rwy
GET https://markdownregistry.com/api/v1/resolve?ref=sshadows/lethal/coverage-differential
GET https://markdownregistry.com/api/v1/blob/7a53074abdcc3c855f92dbaa375f1869a731a551070bc371f1b79f95c1af182b

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from sshadows/lethal

al-compile SKILL.md skill
sshadows/lethal · .claude/skills/al-compile/SKILL.md · Offline-compile an AL project with alc.exe and report diagnostics — the standalone verification for AL edits (AL has no…
git:20261021.85ce5d5 · audit A · 10 stars
al-probe SKILL.md skill
sshadows/lethal · .claude/skills/al-probe/SKILL.md · Scaffold, publish and run a throwaway AL probe app against a LethAL BC container to MEASURE real Business Central…
git:20261106.09fab17 · audit A · 10 stars
control-app SKILL.md skill
sshadows/lethal · .claude/skills/control-app/SKILL.md · Rebuild the LethAL Control BC extension and publish it to the target containers. Use after editing anything under…
git:20261021.f929b5a · audit A · 10 stars
live-gate SKILL.md skill
sshadows/lethal · .claude/skills/live-gate/SKILL.md · Run the LethAL live integration gate against the Cronus BC container — the frozen per-mutant tables through both…
git:20261023.f924fd4 · audit A · 10 stars
measurement-campaign SKILL.md skill
sshadows/lethal · .claude/skills/measurement-campaign/SKILL.md · Use when running LethAL against a real project or customer app, measuring a real codebase's suite, comparing two live…
git:20261021.98334d7 · audit A · 10 stars
recover-tier SKILL.md skill
sshadows/lethal · .claude/skills/recover-tier/SKILL.md · Recover a LethAL tier that is quarantined or holding a stranded operation marker — recycle the environment, force-reset…
git:20261003.fa4847c · audit A · 10 stars
release SKILL.md skill
sshadows/lethal · .claude/skills/release/SKILL.md · Cut a LethAL release end to end — version bump, CHANGELOG section, README version references, the local and live gates…
git:20261021.98334d7 · audit A · 10 stars
repo-scripts SKILL.md skill
sshadows/lethal · .claude/skills/repo-scripts/SKILL.md · Use BEFORE writing an inline shell or bun one-off for any of these: running the build/test loop (typecheck, clean dist…
git:20261022.3fc2385 · audit A · 10 stars
CLAUDE.md CLAUDE.md file
sshadows/lethal · CLAUDE.md
git:20261106.68688a2 · audit A · 10 stars
lethal-mutation-testing SKILL.md skill
sshadows/lethal · skills/lethal-mutation-testing/SKILL.md · Run LethAL mutation testing against a Business Central AL project and read the result. Use when asked whether an AL…
git:20261106.0937045 · audit A · 10 stars

Every file in sshadows/lethal

Browse by kind, by grade A, or by owner.