Skip to content
Development
Skill

/gate-build

Runs configured build and test commands in tiered order. Captures output as evidence per tier. Returns PASS, FAIL, WARN, or SKIP based on per-tier verdicts. Internal gate — invoked by verify, not directly by users.

From plugin
specwright
923 skills9 agents18 commands
Install
$ npx -y skills add Obsidian-Owl/specwright --skill gate-build --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/gate-build

Context preview

The summary Claude sees to decide when to auto-load this skill.

Runs configured build and test commands in tiered order. Captures output as evidence per tier. Returns PASS, FAIL, WARN, or SKIP based on per-tier verdicts. Internal gate — invoked by verify, not directly by users.

SKILL.md

gate-build.SKILL.md
name: gate-build
description: >-
  Runs configured build and test commands in tiered order. Captures output as
  evidence per tier. Returns PASS, FAIL, WARN, or SKIP based on per-tier
  verdicts. Internal gate — invoked by verify, not directly by users.
allowed-tools:
  - Read
  - Bash
  - Glob
  - Write

Gate: Build

Goal

Confirm the codebase compiles and tests pass across all configured tiers. This is the most basic gate — if the code doesn't build or tests don't pass, nothing else matters.

Inputs

  • `{projectArtifactsRoot}/config.json` -- `commands.build`, `commands.test`,

`commands.test:integration`, `commands.test:smoke`

  • `{repoStateRoot}/work/{selectedWork.id}/workflow.json` -- selected work unit for evidence path

Outputs

  • Evidence file at `{workDir}/evidence/build-report.md`
  • Gate status update in the selected work's `workflow.json`: PASS, FAIL, WARN, or SKIP
  • Console output showing results inline (users see findings, not just badges)

Constraints

**Tiered Execution (LOW freedom):**

Run tiers in execution order: `commands.build` → `commands.test` → `commands.test:integration` → `commands.test:smoke`. Each tier runs only if the previous tier passes (any tier failure stops further execution).

  • If a tier's command is null or not configured, that tier produces SKIP. Emit

an INFO note for each unconfigured tier — visibility rule: display the unconfigured tier name so users know what was skipped.

  • If all tier commands are null/unconfigured, gate status is SKIP.
  • Timeout: 5 minutes per command. If exceeded, status is ERROR.

**Per-Tier Verdicts (LOW freedom):**

| Tier | Failure verdict | |------|----------------| | `commands.build` | FAIL | | `commands.test` | FAIL (test failure = FAIL) | | `commands.test:integration` | FAIL (integration failure = FAIL) | | `commands.test:smoke` | WARN (smoke produces WARN on failure — charter exception: quality gates default to FAIL: smoke tests validate optional end-to-end paths; degraded smoke is advisory, not blocking) |

Unconfigured tier = SKIP.

**Evidence (LOW freedom):**

Follow `protocols/evidence.md`. Evidence starts with a tier layout header summarizing configured tiers and their run order. Each tier produces its own evidence section containing: command executed (capture command run), exit code, stdout/stderr output, and duration (elapsed wall time). Per-tier sections are written regardless of tier outcome.

  • Follow `protocols/evidence.md#verdict-rendering` for verdict rendering.
  • Update the selected work's `workflow.json` gates section per `protocols/state.md`.

Protocol References

  • `protocols/evidence.md#verdict-rendering` -- default-FAIL, self-critique, visibility
  • `protocols/evidence.md` -- evidence storage and freshness
  • `protocols/state.md` -- gate status updates

Failure Modes

| Condition | Action | |-----------|--------| | Build command not configured | SKIP build check, continue to test check | | Test command not configured | SKIP test check | | test:integration command not configured | SKIP integration tier, emit INFO note | | test:smoke command not configured | SKIP smoke tier, emit INFO note | | All tier commands null | Gate status = SKIP | | Command times out (>5min) | Gate status = ERROR with timeout message | | test:integration failure | FAIL verdict (integration failures block ship) | | test:smoke failure | WARN verdict (smoke is advisory; see charter exception: quality gates default to FAIL) |

Read more
Ships withspecwright

Craft quality software with AI discipline. Spec-driven development plugin for Claude Code and Opencode — quality gates, adversarial testing, and evidence capture.

Get the whole plugin

Other skills on specwright.