Merge reviewed PR #3838

This commit is contained in:
haelyra committed 2026-11-05 23:11:09 -05:00
commit acb2ac57a4
13 files changed
+37 -13

No files matched your search

+1
View File
@@ -4,6 +4,7 @@
### Changed
- Rename the `claude-devfleet` skill to `devfleet` so its frontmatter name complies with Claude's reserved-word rule; the `/devfleet` entrypoint and external DevFleet integration are unchanged (#3783).
- GateGuard first-touch denials ask questions that fit the target: by class (instruction, test, prose, config, code) on its project-relative path, and for code by what the change touches. An edit that changes no public declaration asks for the in-module call sites instead of importers and public API, and the data-schema question is dropped when the change handles no data. Changes the hook cannot read with confidence keep the original four questions word for word.
- GateGuard's external-contract question ("name what outside this repository decides the format, units, timezone or protocol semantics") is opt-in through `GATEGUARD_EXTERNAL_CONTRACT_QUESTION`. Hidden-intent trials did not show it helps, so by default a code change that handles data asks only the data-schema question, and an unreadable change gets the original four questions and condensed hint.
- GateGuard credits investigation that already happened: a completed, non-error `Glob`, `Grep`, `LS`, or search-style Bash/PowerShell command in the current human turn that names the target and whose scope contains it passes the first touch with a note. `Read` never counts, nor do searches in the same batch, before a compaction, with error results, or whose filters exclude the target or limit the search to another directory. When a denial follows a search that nearly counted, it names that search and why it did not count.
+1 -1
View File
@@ -264,7 +264,7 @@ These slash commands were retired in favor of skills. The command files still ex
| `/docs` | `documentation-lookup` |
| `/claw` | `nanoclaw-repl` |
| `/context-budget` | `context-budget` |
| `/devfleet` | `claude-devfleet` |
| `/devfleet` | `devfleet` |
| `/orchestrate` | `dmux-workflows` and `autonomous-agent-harness` |
| `/prompt-optimize` | `prompt-optimizer` |
| `/rules-distill` | `rules-distill` |
+1 -1
View File
@@ -33,7 +33,7 @@ skills:
- carrier-relationship-management
- change-impact
- ck
- claude-devfleet
- devfleet
- click-path-audit
- clickhouse-io
- brownfield-retrospective
@@ -1,5 +1,5 @@
---
name: claude-devfleet
name: devfleet
description: Claude DevFleet経由でマルチエージェントコーディングタスクをオーケストレーション — プロジェクトを計画し、分離された作業ツリー内で平行エージェントを派遣し、進捗を監視し、構造化レポートを読む。
origin: community
---
@@ -1,5 +1,5 @@
---
name: claude-devfleet
name: devfleet
description: 通过Claude DevFleet协调多智能体编码任务——规划项目、在隔离的工作树中并行调度智能体、监控进度并读取结构化报告。
origin: community
---
+4 -4
View File
@@ -1,14 +1,14 @@
---
description: Legacy slash-entry shim for the claude-devfleet skill. Prefer the skill directly.
description: Legacy slash-entry shim for the devfleet skill. Prefer the skill directly.
---
# DevFleet (Legacy Shim)
Use this only if you still call `/devfleet`. The maintained workflow lives in `skills/claude-devfleet/SKILL.md`.
Use this only if you still call `/devfleet`. The maintained workflow lives in `skills/devfleet/SKILL.md`.
## Canonical Surface
- Prefer the `claude-devfleet` skill directly.
- Prefer the `devfleet` skill directly.
- Keep this file only as a compatibility entry point while command-first usage is retired.
## Arguments
@@ -17,7 +17,7 @@ Use this only if you still call `/devfleet`. The maintained workflow lives in `s
## Delegation
Apply the `claude-devfleet` skill.
Apply the `devfleet` skill.
- Plan from the user's description, show the DAG, and get approval before dispatch unless the user already said to proceed.
- Prefer polling status over blocking waits for long missions.
- Report mission IDs, files changed, failures, and next steps from structured mission reports.
+1 -1
View File
@@ -963,7 +963,7 @@
"skills/ai-first-engineering",
"skills/autonomous-loops",
"skills/blueprint",
"skills/claude-devfleet",
"skills/devfleet",
"skills/content-hash-cache-pattern",
"skills/continuous-agent-loop",
"skills/cost-aware-llm-pipeline",
+1 -1
View File
@@ -188,7 +188,7 @@
"carrier-relationship-management": "supply chain/logistics domain pack",
"cisco-ios-patterns": "network-device ops niche domain",
"ck": "Claude-Code-only persistent memory commands",
"claude-devfleet": "multi-agent orchestration via external DevFleet product",
"devfleet": "multi-agent orchestration via external DevFleet product",
"codehealth-mcp": "wraps CodeScene MCP SaaS",
"competitive-platform-analysis": "competitive/marketing analysis",
"competitive-report-structure": "competitive/marketing reporting",
+1 -1
View File
@@ -198,7 +198,7 @@
"skills/brand-voice/",
"skills/carrier-relationship-management/",
"skills/change-impact/",
"skills/claude-devfleet/",
"skills/devfleet/",
"skills/cisco-ios-patterns/",
"skills/clickhouse-io/",
"skills/code-tour/",
+1 -1
View File
@@ -36,7 +36,7 @@ Exclude anything that:
| `carrier-relationship-management` | supply chain/logistics domain pack |
| `cisco-ios-patterns` | network-device ops niche domain |
| `ck` | Claude-Code-only persistent memory commands |
| `claude-devfleet` | multi-agent orchestration via external DevFleet product |
| `devfleet` | multi-agent orchestration via external DevFleet product |
| `codehealth-mcp` | wraps CodeScene MCP SaaS |
| `competitive-platform-analysis` | competitive/marketing analysis |
| `competitive-report-structure` | competitive/marketing reporting |
+9
View File
@@ -11,6 +11,8 @@
* 4. SKILL.md frontmatter `description:` uses an inline scalar — not a
* literal block scalar (`|` / `|-` / `|+`), which preserves internal
* newlines and breaks flat-table renderers keyed off `description`.
* 5. SKILL.md frontmatter `name:` does not contain the reserved words
* `claude` or `anthropic`.
*
* Frontmatter findings default to WARN so CI does not break while
* pre-existing data defects are being cleaned up out of band (see #1663).
@@ -264,6 +266,13 @@ function validateSkillFile(skillMd, label, reportFrontmatterFinding, opts = {})
reportFrontmatterFinding(`${label} - frontmatter missing required field: name`);
} else if (values.name === '') {
reportFrontmatterFinding(`${label} - frontmatter 'name' is empty`);
} else {
const reservedName = /(?:anthropic|claude)/i.exec(values.name);
if (reservedName) {
reportFrontmatterFinding(
`${label} - frontmatter 'name' contains reserved word '${reservedName[0]}'`
);
}
}
if (!Object.prototype.hasOwnProperty.call(values, 'description')) {
@@ -1,5 +1,5 @@
---
name: claude-devfleet
name: devfleet
description: Orchestrate multi-agent coding tasks via Claude DevFleet — plan projects, dispatch parallel agents in isolated worktrees, monitor progress, and read structured reports. Use when dispatching parallel coding agents across isolated worktrees and tracking their reports.
metadata:
origin: community
+14
View File
@@ -1200,6 +1200,20 @@ function runTests() {
cleanupTestDir(testDir);
})) passed++; else failed++;
if (test('errors when a skill name contains a reserved word (strict mode)', () => {
const testDir = createTestDir();
const skillDir = path.join(testDir, 'claude-devfleet');
fs.mkdirSync(skillDir);
fs.writeFileSync(path.join(skillDir, 'SKILL.md'),
'---\nname: claude-devfleet\ndescription: "Multi-agent orchestration."\n---\n# Skill');
const result = runSkillsValidator(testDir, ['--strict']);
assert.strictEqual(result.code, 1, '--strict must reject reserved skill names');
assert.ok(result.stderr.includes("frontmatter 'name' contains reserved word"),
`Should report the reserved skill name; got stderr: ${result.stderr}`);
cleanupTestDir(testDir);
})) passed++; else failed++;
if (test('warns on literal block-scalar description (|-)', () => {
const testDir = createTestDir();
const skillDir = path.join(testDir, 'block-desc-skill');