Joel Stevick (gmail)
Community@AiFirstDevelopment
AiFirstDevelopment provides a quorum-based change delivery system covering planning, building, multi-lens review, adjudication, regression testing, and CI gating for repositories.
Agent Skills by Joel Stevick (gmail)
Showing 10 vetted skills indexed across 1 GitHub repositories.
AiFirstDevelopment3-review
Reviews a built change from six independent lenses and writes one review file per lens.
AiFirstDevelopmentpipeline
Runs an autonomous build, six-lens review, and adjudication pipeline from an approved plan.
AiFirstDevelopmentstatus
Reports pipeline progress for a work item and names the next command to run.
AiFirstDevelopment4-quorum
Adjudicates code review findings, applies accepted fixes, and writes a verdict file.
AiFirstDevelopmentaudit
Measures a repository against a spec and reports gaps without writing code.
AiFirstDevelopmenthistory
Lists every change planned through the quorum pipeline from docs/work/ and git history.
AiFirstDevelopment2-build
Implements an approved plan file step by step, recording deviations and test results.
AiFirstDevelopmentcalibrate
Measures a six-lens code review panel's catch rate and false-positive rate against fixtures with planted defects.
AiFirstDevelopment1-plan
Writes falsifiable implementation plans with acceptance criteria to docs/work/<slug>/plan.md before any code exists.
AiFirstDevelopmentguard
Runs mechanical integrity checks on pipeline artifacts and installs them as a CI gate.
Frequently Asked Questions About Joel Stevick (gmail)
FAQPage SchemaWhat tasks can I accomplish with AiFirstDevelopment's skills?▼
You can plan a change with acceptance criteria, implement it step-by-step, run six independent review lenses, adjudicate findings into a verdict, write and run behavioral regression tests, wire tests into CI as required status checks, audit a repo against a spec, and track pipeline status and history.
Who should use the quorum pipeline skills?▼
Software engineers and technical leads who want structured, evidence-based change delivery. It suits teams needing enforced regression gates, independent multi-perspective code review, and auditable records of plans, reviews, and verdicts stored under docs/work/ in the repository.
How does a typical run of the quorum pipeline work?▼
Start with 1-plan to write intent and acceptance criteria to docs/work/<slug>/plan.md, then 2-build implements it, 3-review writes one file per lens, and 4-quorum adjudicates findings until tests pass. The pipeline skill runs all steps autonomously after plan approval.
Can I check a repository that never used the pipeline?▼
Yes. The audit skill measures any existing repository against a spec supplied as free text or a file path, reporting where the implementation falls short or returning 'All clear'. It runs on the default branch, writes no code, and requires no prior pipeline adoption.
What prerequisites and safeguards do these skills require?▼
Skills operate on a git repository and store artifacts under docs/work/<slug>/. The guard skill verifies requirements are unchanged, no test was weakened, reviews are append-only, and evidence is real; it can install these checks as a CI gate. The calibrate skill measures review-lens catch rates against known-defect fixtures.