Menu
Skills neste repositório
O SkillsMP coletou 64 skills de aiosbrain/aios-workspace. Abra uma skill para revisar a origem e os detalhes.
aiosbrain/aios-workspaceMostrando 24 de 64 skills coletadas.
Review meeting transcripts through the canonical aios transcripts CLI and typed review engine. Stages grounded decisions and explicit task commitments in an owner-local V2 record, reports TD1-TD6 grading, and requires one human approval before either local…
Idioma do texto original: inglês
Set up, update, or incrementally extend the owner's profile in this workspace: interview the user (or draft from a link), then write it into the workspace memory files (.claude/memory/USER.md + WORKSPACE.md) so the agent has durable context. Also handles…
Idioma do texto original: inglês
Pull today's Google Calendar events and unread Gmail inbox threads via the `gog` CLI and write them as normalized comms activity records into 1-inbox/comms/activity.jsonl — the writer the operator loop's comms source (src/operator-loop/sources/comms.ts) reads…
Idioma do texto original: inglês
Verify an agent wrap-up's checkable claims — review AI-generated diffs, PRs, commits, or implementation summaries and validate assertions like "CI green", "secrets rotated", "tests added", or "mergeable" against reality, plus security, correctness, and…
Idioma do texto original: inglês
Produce an AI-transformation maturity report + roadmap deliverable from the three AM scopes — the individual (workstation), the codebase(s), and the team. Pulls the owner's placement, scores the relevant repos, folds in the team rollup if a brain is…
Idioma do texto original: inglês
Produce a weekly team digest (decisions, scope moves, task deltas, risks) with a rubric-gated self-correction loop: an independent verifier grades the draft against .claude/rubrics/weekly-synthesis.md and the harness revises until the rubric passes or its…
Idioma do texto original: inglês
Classify every unmerged remote branch as truly-unmerged, already-shipped-under-a-different-hash (squash-merge duplicate), or genuinely stale — with evidence, not guesswork. Use when the user asks "are these branches actually stale", "check unmerged branches",…
Idioma do texto original: inglês
Check whether gitignored build artifacts (e.g. dist/) that runtime code depends on are present and fresh, in this checkout and every registered worktree. Use before a demo, on any fresh clone/worktree, or when the user asks "before the demo", "pre-demo…
Idioma do texto original: inglês
Find test files that exist on disk but run in neither `npm test` nor CI (orphaned tests = false confidence), plus stale/inaccurate coverage reports. Use when the user asks "is this test actually running", "check test wiring", "orphaned tests", or invokes…
Idioma do texto original: inglês
Use the codebase knowledge graph for structural code queries. Triggers on: explore the codebase, understand the architecture, what functions exist, show me the structure, who calls this function, what does X call, trace the call chain, find callers of, show…
Idioma do texto original: inglês
Collect and push AI provider spend (Cursor + Claude) to the Team Brain. Use when the user asks about token usage, AI costs, billing, overages, or at end-of-day during heavy build sprints. Run before sync to update the team-visible spend ledger.
Idioma do texto original: inglês
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for…
Idioma do texto original: inglês
Known framework gotchas and stale patterns. Fast lookup before writing framework code. Use when creating or modifying files that import from Next.js, Supabase, React, or other tracked frameworks.
Idioma do texto original: inglês
Critically review implementation plans and produce a copy-paste feedback prompt for another model, especially Opus, to adapt the plan. Use when the user invokes review-plan or asks for an adversarial, technical, security, functional, sequencing, or…
Idioma do texto original: inglês
Place the workspace owner on the Agentic Maturity (AM) model and give them a prescribed next step. Seeds objective signals from `aios analyze` (local session logs) when available, confirms with a short interview, applies the verification cap, writes the…
Idioma do texto original: inglês
Read one or a handful of web pages (a personal site, profile, or company page) with Firecrawl and return structured profile facts (person, company, focus areas, tools). Use when the user wants to draft or enrich their workspace profile from a link — e.g.…
Idioma do texto original: inglês
A harmless instructions-only skill for testing the scanner's low classification. It only contains prose guidance and a reputable docs link.
Idioma do texto original: inglês
A skill that bundles a benign helper script. It should classify as elevated — code is present, but there is no network egress, secret read, external URL, or injection signal.
Idioma do texto original: inglês
A hostile fixture for the scanner — bundles exfil code, hides a zero-width char, uses injection phrasing, and links an external host. Must classify as high.
Idioma do texto original: inglês
Test fixture — a skill that bundles an EXTENSIONLESS executable helper (a shebang script with no .py/.sh extension). Verifies the scanner detects bundled code and its high-signal content even when no file extension classifies it. Deliberately for tests.
Idioma do texto original: inglês
Sync this workspace with the AIOS Team Brain. Use when the user says "sync", "push to the brain", "pull team updates", or at a natural stopping point after producing deliverables/decisions. Reviews what is blocked and why before pushing — promotion is…
Idioma do texto original: inglês
Answer a question by traversing the local OKF bundle — reading index files to orient, fanning out to relevant documents, following cross-links for context, and synthesizing a cited answer. No brain required; works fully offline. Use when the Team Brain is…
Idioma do texto original: inglês
Governance audit of an engagement decision log. Runs a dynamic multi-agent workflow: one verifier per rule (structurally-earned coverage of every entry) followed by an adversarial verification pass that removes false positives. Use when asked to audit / QA /…
Idioma do texto original: inglês
Detect scope creep across an engagement's deliverables. Runs a dynamic multi-agent workflow: classify each deliverable against the scope baseline/ledger, then an adversarial reviewer re-grades each flag to the lowest defensible severity (removing false…
Idioma do texto original: inglês