Back to skills
SKILL.md
Requests
ASecurity> **Created**: 2026-04-08 > **Status**: In Progress > **Priority**: P2 > **Tech Spec**: [2-tech-spec.md](../2-tech-spec.md)
- 192 stars
- 0 votes
- 0 copies
- 1 view
- Added September 4, 2026
Works with
Security analysis
100/100Pro scans all 2 files and shows the line behind each finding
npx -y skills add sd0xdev/sd0x-harness --skill requests --agent claude-codeAre you the author of Requests?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/sd0xdev-requests-41162620)# Context-Aware Q&A Skill `/ask`
> **Created**: 2026-04-08
> **Status**: In Progress
> **Priority**: P2
> **Tech Spec**: [2-tech-spec.md](../2-tech-spec.md)
## Background
開發者在 sd0x-dev-flow 開發過程中,需要快速了解 codebase、git history、rules、docs、skills 等上下文資訊。現有 skill 要麼太重(`/deep-research`)要麼太專門(`/code-explore`、`/codex-explain`)。需要一個輕量級、通用型的 Q&A skill,能自動收集 context 並結合 conversation history 深度回答問題。
## Requirements
- 使用者透過 `/ask <question>` 觸發
- 自動偵測 session context(branch、changed files、feature、recent commits)
- LLM-inferred 問題意圖分類(code / git / docs / rules / skill / arch / multi)
- 根據意圖執行 per-intent context gathering pipeline
- 整合 conversation context,提供比直接問更深刻的回答
- 按需調用 sub-agent(max 2, complexity-based)
- 當問題更適合其他 skill 時,主動路由建議
- 嚴格 read-only,不觸發 auto-loop
## Scope
| Scope | Description |
|-------|-------------|
| In | SKILL.md + references + tests + catalog integration + CLAUDE.md / `.claude/CLAUDE.md` / `CLAUDE.template.md` 三檔更新 + README catalog 更新 |
| Out | 外部 API 整合、web research、code 修改功能 |
## Related Files
| File | Action | Description |
|------|--------|-------------|
| `skills/ask/SKILL.md` | New | Main skill definition(~120 lines) |
| `skills/ask/references/intent-patterns.md` | New | Intent classification 範例與邊界案例 |
| `skills/ask/references/routing-table.md` | New | Skill routing decision table |
| `test/skills/ask.test.js` | New | Skill lint + routing + provenance + read-only + path security tests |
| `docs/skill-catalog.yml` | Modify | Add `/ask` entry |
| `CLAUDE.md` | Modify | Add `/ask` to command table |
| `.claude/CLAUDE.md` | Modify | Add `/ask` to command table |
| `CLAUDE.template.md` | Modify | Add `/ask` to command table(3-file parity) |
## Acceptance Criteria
- [x] SKILL.md 通過 `skill-lint.js` 所有 P0/P1/P2 checks
- [x] Intent classification 涵蓋 7 個 intent types,每個有 per-intent context gathering pipeline
- [x] Sub-agent dispatch 基於 complexity(simple: 0 agents、medium: 1、complex: 2 max)
- [x] Skill routing:支援至少 8 個 route target(`/feature-dev`、`/codex-review-fast`、`/review-spec`、`/codex-review-doc`、`/bug-fix`、`/next-step`、`/deep-research`、`/code-explore`),路由建議含理由
- [x] Source evidence 支援 3 種 type(file / commit / command),output 含 Sources table
- [x] Read-only + path security:SKILL.md 含 prohibited git commands 清單、`allowed-tools` 無 Edit/Write;repo boundary enforcement + `.env`/`credentials.*`/`*secret*` skip patterns + output secret redaction
- [ ] Conversation context integration:reuse 先前 conversation 中的 active feature/file context 來豐富問題理解;當前 turn 有相關 prior context 時,Sources 中引用之
- [ ] Doc discovery 使用 feature-first 策略(`canonical_docs` → fallback glob)
- [ ] Pass `/codex-review-fast`
- [ ] Pass `/precommit-fast`
## Progress
| Phase | Status | Note |
|-------|--------|------|
| Analysis | Done | req-analyze + tech-spec completed |
| Development | In Progress | Implementation identified heuristically by batch `--update-all` (2026-08-15). This ticket records no per-AC `file:line` evidence — producing that is what `--verify-ac` is for |
| Testing | - | |
| Acceptance | In Progress | 6/10 AC verified against the repo by batch `--update-all` (2026-08-15); closure-grade sign-off still needs `--verify-ac` |
## References
- Tech Spec: [2-tech-spec.md](../2-tech-spec.md)
- Feature Context Resolution: [feature-context-resolution.md](../../../../skills/create-request/references/feature-context-resolution.md) — repointed 2026-08-11: the `/tech-spec` duplicate this line named was merged into this canonical copy
Attribution
Comments
Loading comments…