Om Setup Agent Pipeline

Setup Agent Pipeline

Every skill in this collection reads its repository-specific settings from .ai/agentic.config.json. This skill writes that file. It is the first skill to run in a fresh repository; the others stop and point here when the config is missing.

Arguments

• --defaults (optional) — skip all questions and write the auto-detected config without confirmation.

Config schema

.ai/agentic.config.json, committed to the repository:

{
  "version": 1,
  "baseBranch": "auto",
  "tracker": "github",
  "browser": { "provider": "agent-browser" },
  "validation": {
    "commands": ["pnpm typecheck", "pnpm test", "pnpm build"]
  },
  "labels": {
    "enabled": true,
    "pipeline": ["review", "changes-requested", "qa", "qa-failed", "merge-queue", "blocked", "do-not-merge"],
    "category": ["bug", "feature", "refactor", "security", "dependencies", "documentation"],
    "meta": ["needs-qa", "skip-qa", "qa-approved", "qa-self-verified", "in-progress", "ci-monitoring"],
    "priority": ["priority-low", "priority-medium", "priority-high", "priority-extreme"],
    "risk": ["risk-low", "risk-medium", "risk-high"]
  },
  "qaGate": true,
  "ci": { "maxWaitMinutes": 40 },
  "engine": { "loopStepThreshold": 20, "executorTier": "standard", "stepReview": "final" },
  "paths": {
    "runs": ".ai/runs",
    "analysis": ".ai/analysis",
    "specs": ".ai/specs",
    "scripts": ".ai/scripts",
    "qa": ".ai/qa"
  },
  "reviewChecklist": null,
  "closeKeywords": []
}

Field reference:

• baseBranch — the branch PRs target. "auto" means resolve at runtime from the repository's default branch; set an explicit name only when PRs target something else.
• tracker — the issue/PR tracker provider. Selects the tracker descriptor at .ai/trackers/<tracker>.md, which defines how every tracker operation the skills name is executed. The collection ships "github" (the gh CLI); other trackers are added by writing one descriptor file — see Tracker providers below.
• browser.provider — the browser-automation provider used by QA and integration-test skills. Selects .ai/browsers/<provider>.md. Fresh setups default to "agent-browser"; configs without this key keep legacy Playwright behavior (see Browser providers).
• validation.commands — ordered list of shell commands that constitute the full validation gate. Skills run them in order and treat any non-zero exit as a gate failure. Keep the list complete: typecheck, lint, tests, build — whatever proves the repo is healthy.
• labels.enabled — when false, skills skip every label operation and note that in their PR summaries. Use this for repos that do not want the label workflow.
• labels.pipeline — mutually exclusive workflow states. A PR carries at most one.
• labels.category — additive kind-of-change labels.
• labels.meta — additive process labels. needs-qa requests manual QA; skip-qa opts out (never combine the two); qa-approved records that QA passed; qa-self-verified marks the self-QA exception; in-progress is the claim lock automated skills apply while they are actively working the item; ci-monitoring says the work is finished and fully reported — labels applied, review submitted, comments posted — and the agent is only watching the CI run, so it is not a claim and another agent or a human may act on the PR freely (it means one thing only: the CI-result follow-up comment is still owed). One label lives outside the config taxonomy: do-not-close, applied by humans to issues that housekeeping skills must never auto-close — skills only ever read it.
• labels.priority — mutually exclusive urgency of the work. Unset is treated as medium.
• labels.risk — mutually exclusive blast radius of the change. Unset is treated as medium. Priority is how urgent the work is; risk is how dangerous the change is to ship.
• qaGate — when true, a PR carrying needs-qa must not merge until it also carries qa-approved, even when every other check is green. When false, needs-qa is advisory only.
• ci.maxWaitMinutes — the hard cap, in minutes, on how long any skill waits for CI to settle before it stops waiting (default 40). It is a safety valve, not a merge gate: when the budget runs out the skill runs the local validation.commands gate as its completion evidence, posts the bail-out comment, drops ci-monitoring, and exits cleanly instead of hanging on a run that may take hours. Raise it for slow pipelines, lower it for fast ones; 0 disables waiting entirely (report immediately and never follow up). Required checks still gate the actual merge no matter what this is set to.
• engine.executorTier — optional; the default abstract model tier (cheap / standard / capable) for executor subagents dispatched by the loop skills when a Tasks-table Exec cell names none. Harnesses that support subagent model selection map the tier onto their closest model class; others ignore it. Configs without the key behave as standard.
• engine.loopStepThreshold — the Step count above which om-auto-create-pr hands a run off to om-auto-create-pr-loop (default 20). Raise it to keep more runs on the cheaper plain engine; --loop always forces the loop regardless.
• engine.stepReview — optional; how often the loop skills code-review landed work mid-run: final (default — only the authoritative end-of-run review), checkpoint (review the diff at every checkpoint pass), or per-step (review each Step's commit as it lands). Blocker/major findings are fixed immediately as X.Y-review-fix Steps; minors defer to the final review, which runs in every mode.
• paths.runs — where execution plans of autonomous runs are stored.
• paths.analysis — where generated reports are stored.
• paths.specs — where feature specifications live (default .ai/specs). Spec filenames follow {YYYY-MM-DD}-{kebab-case-title}.md. om-spec-writing writes here, om-prepare-issue links from here, om-followup-issue-from-pr checks here first in design-doc mode, and om-brainstorm writes handoff briefs under <paths.specs>/briefs/.
• paths.scripts — where reusable environment scripts are generated (default .ai/scripts); om-prepare-test-env writes the env bring-up/teardown scripts here.
• paths.qa — where QA working state and artifacts live (default .ai/qa): the shared test-env.json descriptor, and QA reports/screenshots under <paths.qa>/artifacts_<runId>/.
• reviewChecklist — optional path to a repo-local review checklist file. When set, the om-code-review skill reads it in addition to its built-in checklist. A root CODE_REVIEW.md (see Project docs) is always picked up regardless.
• closeKeywords — optional list of extra words that mark a PR as closing an issue, for repositories whose PR bodies are not written in English. om-close-fixed-issues matches the built-in English keywords (fix/fixes/fixed, close/closes/closed, resolve/resolves/resolved) plus everything listed here, case-insensitively and only immediately before a #N token; configured words extend the built-ins and never replace them. The tracker's own closingIssuesReferences parse is English-only too, so a Polish repo writing Zamyka #88 gets no closing signal from either source until it sets, for example, ["zamyka", "naprawia", "rozwiązuje"]. Leave it empty on an English repository. Whatever the setting, a run that finds issue mentions without a recognized keyword reports them rather than passing over them silently.

Tracker providers

No skill in this collection calls a tracker CLI or API directly. Skills name tracker operations — get-issue, create-pr, comment-pr, merge-pr, and the rest of the contract in references/trackers/TEMPLATE.md — and the repository's tracker descriptor at .ai/trackers/<tracker>.md (selected by the tracker config field) defines how each operation is executed. This skill installs the descriptor: it copies the shipped implementation from its own references/trackers/<tracker>.md into the repo, where it is committed alongside the config.

The repo's copy is authoritative, which is also the extension mechanism: teams edit .ai/trackers/<tracker>.md to extend or override any operation, and every skill picks the change up on its next run. A whole new tracker (e.g. Linear) is ONE new descriptor file written from TEMPLATE.md, plus the matching tracker value; split setups (issues in Linear, PRs on GitHub) implement the issue operations against the issue tracker and delegate the PR sections to the GitHub descriptor, as the template describes.

The collection ships github.md; unshipped trackers are scaffolded from references/trackers/TEMPLATE.md (see step 4 and Rules).

Browser providers

Browser-capable skills use the same committed-descriptor pattern as trackers: they name provider operations (ensure-installed, doctor, open, snapshot, interact, assert, screenshot, close) and read .ai/browsers/<provider>.md, selected by browser.provider. The collection ships agent-browser.md (the self-provisioning fresh-setup default, local processes only) and playwright.md, plus references/browsers/TEMPLATE.md for custom providers. A config without browser.provider is read as playwright for backward compatibility. Full operation contract, agent-browser platform support, and the compatibility path: references/browser-providers.md.

Project docs: SDLC.md, AGENTS.md, CODE_REVIEW.md, BACKWARD_COMPATIBILITY.md

Beyond the config, this skill produces the human-readable half of the pipeline: SDLC.md (ticket flow, label state machine, QA gate, claim protocol), AGENTS.md (project overview plus the task-routing table every skill reads), CODE_REVIEW.md (the repo's review rules, auto-applied by om-code-review), and BACKWARD_COMPATIBILITY.md (the protected contract surfaces skills check against). Every document is derived from the current project, never copied, and generated only when missing — an existing file is never touched. Per-document generation guidance: references/project-docs.md.

Per-skill local overrides

Every skill in this collection checks, right after loading the config, for a repo-local extension of the same name at .ai/skills/<skill-name>/SKILL.md. This skill does not create local skills; it only owns the convention. Full contract — extension semantics, what local rules can and cannot override, the safety clause: references/agentic-setup.md.