← All creators

assaio

Organization

Is your AI coding spend delivering? Offline-first analytics for Claude Code, Codex CLI, Gemini CLI, GitHub Copilot CLI and Cline — $/100 AI lines per project, every verdict carrying its own confidence, a self-contained dashboard, exec plugins in any language, and a self-hosted team server. No telemetry; prompts are never read.

16 indexed · 1 Featured · 63 stars · avg score 86
Prolific

Categories

Indexed Skills (16)

Code & Development Featured

auto

Autonomous, non-interactive run of the whole chain — work → plan → build → gate → review → ship up to the PR — for a task or B-id, without questions. Takes the most reversible decision when unsure and records it in the work file as "decided alone … reverse by …". Stops only at hard lines (merge, tag, push to main, money, protected contracts). Exits cleanly when nothing is left to do.

63 Updated today
assaio
Code & Development Solid

add-source

Checklist for adding a data source (a new parser under internal/parser/) or a metric validator (internal/analyze/) with every trap this repo has hit — the same-commit surfaces, the fuzz and golden contract, the capability matrix, the published counts. Use "new parser", "support tool X", "add a validator", "new metric", "connector".

63 Updated today
assaio
Code & Development Solid

docs

Improves and simplifies this repo's documentation without losing a fact — finds contradictions between surfaces, duplicated facts, sentences no longer true of the binary, pages nobody links, and prose a reader cannot act on; proposes the smaller true version through the second-family door. Use "simplify the docs", "is the README still true", "docs review", "too long", "where should this go".

63 Updated today
assaio
Code & Development Solid

plan

Writes the ONE work file for a task — docs/work/<date>-<slug>.md holding spec, plan, journal and handoff together — with tracks a coder can take unread. Use for M and L work, "plan this", "write the spec", "how would we build". For L work it adds alternatives and a challenger review before any code exists.

63 Updated today
assaio
Code & Development Solid

research

Research for the roadmap — a market or evidence question, a competitor's surface, a milestone's exit criterion, whether a backlog pool is still justified — producing a dated, sourced note under docs/work/ and concrete proposals (ROADMAP evidence rows, BACKLOG items with ids, gates). Use "research", "what does the evidence say", "should we build", "is this still worth it", "roadmap".

63 Updated today
assaio
Code & Development Solid

ship

Ships finished work the way this repo accepts it — one signed Conventional Commit of explicit paths, the changelog/FEATURES/BACKLOG lifecycle, a PR, the deploy check when the merge would publish the site, the work file closed, one retro line. Use "commit this", "open the PR", "ship it", "merge". Never pushes or merges without a yes.

63 Updated today
assaio
AI & Automation Solid

work

Entry point for any task in this repo — "let's do X", "pick up B123", "fix", "add", "change". Checks the ask is still true and unclaimed, sizes it S/M/L, picks the lightest safe path (direct edit, /plan, /research) and says who does what on which model. Use before touching code; skip only for a one-line answer.

63 Updated today
assaio
AI & Automation Solid

build

Executes the tracks of a work file (docs/work/*.md) — in this session or through parallel coder agents when file sets are disjoint — checking every return with a diff and a re-run, a checkpoint per track and a full gate every third. Use after /plan, "build it", "implement the tracks", "continue the work file".

63 Updated today
assaio
AI & Automation Solid

content-model

When and how text goes through the GPT/Gemini door (.claude/scripts/text_model.py) instead of being written by the engineering model — README and docs prose, i18n and help strings, changelog wording, site copy, and any audit or verdict about such text; also the second-family challenge of a research note. Use "rewrite this sentence", "review the wording", "simplify the docs", "second opinion".

63 Updated today
assaio
Code & Development Solid

debug

Root-causes a wrong figure, a failing test, a parser drift or a crash in this repo with the traps that have bitten here — double counting, wrong denominators, windows one day too wide, golden files that hid a bug, a store already migrated. Use "why is this number wrong", "test fails", "doctor reports drift", "reproduce this".

63 Updated today
assaio
Code & Development Solid

gate

Runs this repo's local gate quietly — fmt, lint, test, build, docs regeneration, the harness checks, fuzz when a parser changed, vuln with full — one line per step, the tool's verdict verbatim, log tail only when red. Use before a commit, after a build track, "run the gate", "is it green", "check everything".

63 Updated today
assaio
Code & Development Solid

refactor

Behaviour-preserving restructuring under this repo's norms — split a file past ~200 lines or doing two things, remove dead code with evidence, untangle a dependency pointing outward — with golden files and the corpus as the proof nothing moved. Use "refactor", "split this file", "this is dead code", "simplify". Deleting code is proposed with evidence, not done unasked.

63 Updated today
assaio
Code & Development Solid

release

Prepares the next release per RELEASING.md — green gate, LiteLLM price-table refresh, the public-surface audit through surface-auditor, CITATION.cff, the changelog retitle commit — and stops at the exact tag and push commands for the maintainer. Use "release", "cut a version", "prepare vX.Y.Z", "tag".

63 Updated today
assaio
Code & Development Solid

review

Reviews a change (or the whole repo, or a path) in fresh contexts with this repo's four reviewers — go-reviewer, honesty-auditor, surface-auditor, store-steward when a write path is touched — then runs the review → fix → gate → re-review loop over the fixes only, at most three rounds. Use "review this", "self review", before /ship and before a tag.

63 Updated today
assaio
Code & Development Solid

smoke

Boots the local stack the way a first-time user meets it — build the binary, demo, init on a throwaway store, doctor --strict, dashboard, evidence on this repo, and the team server round trip — and reports what a reader sees. Use "does it still run", "smoke test", "first-run path", "try the binary", after any CLI or ingest change.

63 Updated today
assaio
AI & Automation Solid

ui-check

Checks the offline dashboard and the site in a real browser through the browser agent — every section rendered, empty and unmeasured states as dashes, zero console and network activity, 400 px, keyboard, both themes — with screenshots kept as evidence outside the main context. Use after any dashboard, i18n, validator-text or site change, "check the dashboard", "look at the page".

63 Updated today
assaio

Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.