11 Mile Labs Skills

A multi-harness library of open-source AI agent skills for engineering, QA, product, and work-management workflows, published by 11 Mile Labs.

Published by 11 Mile Labs and presented here as public evidence of AB Whittington's agent-workflow and developer-tooling work: reusable skills, explicit compatibility boundaries, and source you can inspect.

Public source liveGitHub repository

Why it exists

Generic agent craft, separated from business glue.

The repository collects reusable AI coding craft — small skills that work across harnesses — and keeps it deliberately distinct from any vertical business pack, installer, or harness-specific plugin runtime.

It exists so the same craft can be reviewed, copied, and adapted outside the originating repo without dragging private process, project-local pipeline glue, or proprietary tooling along with it.

Design decisions

Skills only, multi-harness, no lock-in.

Each entry ships as a standard SKILL.md and targets Claude Code, OpenAI Codex, and Gemini CLI through shared frontmatter rather than a single runtime.

Originally authored by 11 Mile Labs and adapted from internal tooling, with no copied third-party upstream content — the repository's own README sets that boundary.

Published under MIT with a public contribution guide so any individual skill can be inspected, forked, or replaced without negotiating a license.

Scope boundaries

What was deliberately held back.

Stack-specific procedural packs (Neon, Drizzle, Stripe, Ink, and similar) are deferred until they can ship as coherent skill sets rather than a loose miscellaneous bucket.

Installation is owned by the skills ecosystem; this repository ships content only, with no bespoke installer or harness-specific runtime requirement.

What is inside

Skill catalog.

Every entry below links to its SKILL.md on GitHub. Categories and counts are derived from the typed catalog used to render this page.

Engineering

Skills for sizing change, defending releases, and keeping engineering work scoped.

dependency-upgrade-risk-checkClassify dependency upgrade risk and define proof before updating.
deprecation-plan-briefPlan safe removals with compatibility, notice, proof, and rollback.
implementation-drift-checkCheck changed artifacts against stated intent to catch scope drift and unrelated edits.
regression-risk-mapMap likely regressions and targeted checks for a specific change.
release-readiness-briefDecide whether a change is ready to release, needs evidence, or should be held.
scope-checkAudit scope creep, hidden complexity, and v1/v2 boundaries.

QA

Skills for turning failures and evidence requests into focused work.

bug-repro-briefTurn messy bug reports into reproducible bug briefs before debugging or ticketing.
bugfix-retrospective-noteCapture bugfix cause, proof, prevention, and follow-up without ceremony.
test-failure-signal-briefClassify failing test and check output before debugging, fixing, or rewriting assertions.
test-strategy-slicerChoose focused proof layers for a change instead of defaulting to coverage theater.
verification-contractDefine the evidence required before claiming a feature, ticket, or plan is done.

Product

Skills for tightening requirements, decisions, and pre-commitment thinking.

acceptance-criteria-tightenerRewrite vague acceptance criteria into observable pass/fail statements.
decision-record-snapshotCapture lightweight decision records with source, rationale, scope, and revisit triggers.
devil-advocateArgue against a plan in good faith before committing.
explain-like-5Force a simple explanation of a concept, architecture, or decision.
pre-mortemImagine a project failed and work backward to surface risks.

Work management

Skills for keeping in-flight work transferable, reviewable, and slice-sized.

handoff-readiness-briefMake active work state transferable before pausing, resuming, reviewing, or handing off.
review-comment-action-planTurn review comments into an ordered action plan with clarifications, deferrals, and verification.
ticket-readiness-auditCheck whether a ticket is ready to claim, clarify, split, block, or send back.
work-slice-splitterSplit broad work into independently claimable outcome slices before execution.

Compatibility

Harnesses targeted by the repository.

Harness compatibility per item type, as stated by the repository.
Item	Claude Code	OpenAI Codex	Gemini CLI
Skills	Yes	Yes	Yes
Project memory	CLAUDE.md → AGENTS.md	AGENTS.md	GEMINI.md → AGENTS.md

Evidence trail

Source links.

GitHub repositorySource, history, and repository context.
READMERepository overview, catalog table, and philosophy.
CONTRIBUTINGContribution model and skill-authoring boundaries.
LICENSEMIT license terms.
AGENTS.mdRepository instructions surfaced to AI harnesses.