Automated summary
Operational guide for GitHub work in the Calculadora-BMC repository: branches, commits, pull requests, CI workflows, issue triage, and merge policy.
Strengths
- States a concrete Definition of Done for PRs (green gate:local, filled template, mandatory draft above 500 added lines)
- Actionable red-CI playbook: locate the failing job, pull failed-only logs, known fixes (ALSA headers for native midi, BMC_DISK_PRECHECK_SKIP) and a stop rule after repeated re-kicks
- Clarifies which tooling is available per context (GitHub MCP tools in remote sessions, gh CLI locally, never raw REST calls with tokens from .env)
- Structures triage around priority labels (P0/P1) and requires duplicate search before opening an issue
Limitations
- Tightly coupled to matiasportugau-ui/Calculadora-BMC: job names, npm scripts and workflows do not transfer to other repositories
- Points to internal docs (docs/team/AGILE.md, PROJECT-STATE.md, BACKLOG.md) whose content is not included
- Does not cover authoring or editing GitHub Actions workflows, only running and diagnosing them
Use it whenever a GitHub task targets this repository: creating or reviewing a PR, debugging red CI, managing branches or releases, or triaging issues.
Skip it for generic GitHub questions or for a different repository, since the conventions, workflows and scripts described are specific to Calculadora-BMC.
Security analysis
SafeThe skill is an operational guide for GitHub workflows in a specific repository. It references legitimate development tools (gh CLI, GitHub MCP, npm scripts) and includes safety guardrails such as not exposing secrets, not force-pushing main, and forbidding destructive npm audit fixes without approval. It does not instruct exfiltration, destructive commands, or obfuscated payloads.
No concerns found
How this analysis is madeExamples
My branch fix(deps) has changes ready on Calculadora-BMC. Run through the PR checklist: verify npm run gate:local passes, fill the PR template, and tell me whether the diff size forces draft status.CI is failing on my PR in Calculadora-BMC. Pull the failed job logs, tell me which job broke (validate, lint, env-drift…), and give me the local command to reproduce it.Triage the open issues on the BMC Dev board: check for duplicates, surface priority:P0 and priority:P1 first, and propose labels for each.name: github-expert description: GitHub operations expert for the Calculadora-BMC repo — branches, commits, PRs, CI workflows, review flow, labels/board, and merge policy. Use when creating or reviewing PRs, diagnosing red CI, managing branches/releases, triaging issues, or when the user invokes /github-expert or asks "how do we do X on GitHub" in this repo.
GitHub Expert — Calculadora BMC
Operational guide for all GitHub work on matiasportugau-ui/Calculadora-BMC (default branch: main).
Tooling
- Claude Code on the web / remote sessions: use the GitHub MCP tools (
mcp__github__*) for PRs, comments, CI status, and logs.ghCLI is NOT available there. - Local sessions:
ghCLI works if installed; MCP tools are the fallback. - Never call the GitHub REST API directly with tokens from
.env.
Branch & commit conventions
- Branch names:
feat/...,fix/...,chore/...,docs/...(or the session-designatedclaude/*branch). Never commit directly tomain. - Commit messages: concise, English,
type:prefix —feat,fix,refactor,docs,chore. Scope in parentheses is common:feat(ci): ...,fix(deps): .... - If a branch's PR was already merged, never stack new commits on it — restart the branch from
origin/mainand open a new PR.
PR workflow (Definition of Done — docs/team/AGILE.md §3)
npm run gate:localgreen (lint + test + test:api). For large UI/build changes:npm run gate:local:full(adds build).- Fill the PR template (
.github/PULL_REQUEST_TEMPLATE.md): Qué cambia,Closes #, Tipo checkbox, DoD checklist, Notas para el reviewer. - PR > 500 LOC adds → DRAFT obligatorio and split into atomic commits before marking ready (branch protection can't enforce this; we do).
- If the change alters behavior, append a line under "Cambios recientes" in
docs/team/PROJECT-STATE.mdin the same PR. - Bot/agent-created PRs always start as draft.
CI — what runs and when
ci.yml ("CI — Panelin Calculadora BMC") on push + PR:
| Job | What it does |
|-----|--------------|
| validate | npm test (offline suite) + build |
| lint | ESLint on src/ |
| env-drift | .env.example vs config drift check |
| smoke | Prod smoke (push to main only; MATRIZ CSV is the critical check) |
| channels_pipeline / voice_health / knowledge_antenna | Channel + voice + KB health checks |
Other notable workflows:
deploy-calc-api.yml/deploy-vercel.yml— Cloud Run (panelin-calc, us-central1) and Vercel deploys.catalog-diff.yml— Catalog ↔ MATRIZ diff; upserts a PR comment, fails on regression/S1.codeql.yml,dependency-review.yml— security gates.gemini-*.yml— Gemini review/triage bots (dispatch, review, scheduled triage).smoke-prod-scheduled.yml— scheduled prod smoke; failures open/refresh an auto-managed tracking issue.wa-canonical-cutover.yml— dispatchable WA migrate/flip/soak.labels-sync.yml— label catalogue is code-managed; don't hand-edit labels in the UI.
Red CI playbook
- Identify the failing job (
pull_request_read/actions_get), then pull job logs (get_job_logs,failed_only: true). - Reproduce locally:
npm run gate:local(or the specific script the job runs). - Native
midibuild failures on Linux → missing ALSA headers (libasound2-dev); CI installs them, local containers may not. - Disk precheck false-fails on cloud FS →
BMC_DISK_PRECHECK_SKIP=1. - Fix → commit → push; CI re-runs automatically. Use
actions_run_triggeronly for dispatchable workflows. - Several re-kicks with no progress → stop and report the diagnosis instead of looping.
Issues & board
- Backlog lives in GitHub Issues / Project BMC Dev; offline mirror
docs/team/BACKLOG.md; board rulesdocs/team/AGILE.md. - Priority labels:
priority:P0(drop everything) →priority:P1→ lower. Surface P0/P1 first in any triage. - Search for duplicates before opening an issue; set
state_reasonwhen closing.
Guardrails
npm audit fix --forceis forbidden without explicit approval (it has broken Vite here).- Secrets only in
.env/ Secret Manager — never in code, workflow files, or PR bodies. - Never force-push
main; force-with-lease on feature branches only when rebasing your own unshared work. - Be frugal with bot comments on PRs — comment only when it genuinely adds signal.
Language
Respond in Spanish when the user writes in Spanish. Commands, file paths, and technical terms stay in English.
From the file to a first result
The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.
View the source and supporting filesThis download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.
1. Download the file
Download SKILL.mdFrom the root of a test project, save the file as .skill-downloads/github-expert-for-calculadora-bmc-msh3us9s/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.
Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.
mkdir -p '.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/github-expert-for-calculadora-bmc-msh3us9s/download' --output '.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s/SKILL.md.part' && mv '.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s/SKILL.md.part' '.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s/SKILL.md'2. Check prerequisites
Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.
No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.
The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.
3. Load the skill
After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.
Claude Code
test -f '.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s/SKILL.md' && test -s '.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/github-expert' && cp -R '.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s' '.claude/skills/github-expert'OpenSkills
OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.
Official OpenSkills installationopenskills install './.skill-downloads/github-expert-for-calculadora-bmc-msh3us9s'openskills read 'github-expert-for-calculadora-bmc-msh3us9s'Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.
For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.
openskills syncManual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.
4. Try a first task
Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:
Read .skill-downloads/github-expert-for-calculadora-bmc-msh3us9s/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.
Configure a Chat Channel with SecretRefs
DevOps
Set up chat channels using non-interactive commands, with secrets only as SecretRefs. Covers gather, mutate, repair, and prove steps for reliable delivery.
Investigate ClickHouse CI Failures
DevOps
Investigate a ClickHouse CI failure end-to-end from a PR or S3 report URL. Fetches failed tests, classifies flaky vs real regression, and checks for existing issues/fixes.
Ansible PR Review
DevOps
Reviews an Ansible PR following the project's standardized process, including changelog checks, test coverage validation, and feedback tracking.