Two files, one repository
zhu1090093659/spec_driven_develop ships 2 formats across 2 indexed files. The question worth asking is whether the second one says anything the first does not.
CompareAGENTS.md ↔ CLAUDE.md
| Dimension | Shared | Only in A | Only in B | Overlap |
|---|---|---|---|---|
| Sections | 0 | 5 | 1 | 0% |
| Commands | 0 | 1 | 0 | 0% |
| Section tags | 1 | 1 | 0 | 50% |
What each file covers
Sections
0 shared · 5 only in A · 1 only in B- − Project Agent Instructions
- − Scope
- − Truth Sources
- − Development Rules
- − Validation
- + Claude Code Instructions
Commands
0 shared · 1 only in A · 0 only in B- − git diff --check
Section tags
1 shared · 1 only in A · 0 only in B- − do-not
- agent-behaviour
Line diff
zhu1090093659/spec_driven_develop · AGENTS.md
@@ −1 @@
1# Project Agent Instructions
2
3## Scope
4
5These instructions apply to the whole repository.
6
7## Truth Sources
8
9- Core workflow: `plugins/spec-driven-develop/skills/spec-driven-develop/SKILL.md`
10- Shared workflow references: `plugins/spec-driven-develop/skills/spec-driven-develop/references/`
11- Claude Code sub-agent prompts: `plugins/spec-driven-develop/agents/`
12- OpenCode plugin entrypoint: `plugins/spec-driven-develop/opencode-plugin.js`
13- Plugin manifests: `plugins/spec-driven-develop/.claude-plugin/plugin.json`, `plugins/spec-driven-develop/.codex-plugin/plugin.json`, and the root `.claude-plugin/marketplace.json`
14- Validation and helper scripts: `scripts/`
15- User-facing docs: `README.md` and `README.zh-CN.md`
16- Persistent project memory: use the active coding agent's native project memory surface when available. Do not add a repository fallback memory file unless the workflow explicitly records that choice.
17
18## Development Rules
19
20- Keep this repository Markdown-first. Do not add runtime dependencies unless a requested feature truly needs executable tooling.
21- For workflow behavior changes, update the core `SKILL.md`, the affected references/templates, and the affected agent prompts together. Avoid one-line prompt patches that leave other execution surfaces stale.
22- When changing user-visible behavior, keep `README.md` and `README.zh-CN.md` consistent.
23- `AGENTS.md` and `CLAUDE.md` are active project-level instruction surfaces. Keep them aligned; use `CLAUDE.md` only for Claude Code-specific notes and point back here for shared rules.
24- New features or behavior changes must add or update relevant tests when an automated test surface exists. If no automated suite exists, run the closest static/syntax checks and document why no test was added.
25- Record stable cross-session decisions in the native project memory surface when one is available. If none is available, record only agent-facing rules in `AGENTS.md`/`CLAUDE.md`; do not invent a repo-local memory file without an explicit workflow decision.
26
27## Validation
28
29Run the standing consistency guard first; it subsumes the Python byte-compile check. Then use the closest checks for the changed surface:
30
31- `bash scripts/validate.sh`
32- `bash -n scripts/install-codex.sh scripts/install-cursor.sh scripts/install-opencode.sh scripts/install-agents.sh scripts/install-all.sh`
33- Targeted `rg` checks for newly required workflow language
34- `git diff --check`
35
zhu1090093659/spec_driven_develop · CLAUDE.md
@@ +1 @@
1# Claude Code Instructions
2
3Read `AGENTS.md` first. It is the shared project-level instruction source for Codex, OpenCode, Cursor, Claude Code, and other Markdown-aware coding agents.
4
5Claude Code-specific reminders:
6
7- The optional Claude Code sub-agent prompts live in `plugins/spec-driven-develop/agents/`, including `code-reviewer.md`, the per-lane review-and-fix sub-agent used in the execution loop.
8- The plugin is skills-only: workflows are invoked through `plugins/spec-driven-develop/skills/`, not slash commands.
9- When the shared workflow rules change, update Claude Code-specific prompts only where their execution contract also changes.
10- Do not duplicate shared policy here. Put durable cross-agent rules in `AGENTS.md`; use Claude Code's native project memory surface for stable project facts when available.
11
@@ −1 +1 @@
1−# Project Agent Instructions
1+# Claude Code Instructions
22
3−## Scope
3+Read `AGENTS.md` first. It is the shared project-level instruction source for Codex, OpenCode, Cursor, Claude Code, and other Markdown-aware coding agents.
44
5−These instructions apply to the whole repository.
5+Claude Code-specific reminders:
66
7−## Truth Sources
8−
9−- Core workflow: `plugins/spec-driven-develop/skills/spec-driven-develop/SKILL.md`
10−- Shared workflow references: `plugins/spec-driven-develop/skills/spec-driven-develop/references/`
11−- Claude Code sub-agent prompts: `plugins/spec-driven-develop/agents/`
12−- OpenCode plugin entrypoint: `plugins/spec-driven-develop/opencode-plugin.js`
13−- Plugin manifests: `plugins/spec-driven-develop/.claude-plugin/plugin.json`, `plugins/spec-driven-develop/.codex-plugin/plugin.json`, and the root `.claude-plugin/marketplace.json`
14−- Validation and helper scripts: `scripts/`
15−- User-facing docs: `README.md` and `README.zh-CN.md`
16−- Persistent project memory: use the active coding agent's native project memory surface when available. Do not add a repository fallback memory file unless the workflow explicitly records that choice.
17−
18−## Development Rules
19−
20−- Keep this repository Markdown-first. Do not add runtime dependencies unless a requested feature truly needs executable tooling.
21−- For workflow behavior changes, update the core `SKILL.md`, the affected references/templates, and the affected agent prompts together. Avoid one-line prompt patches that leave other execution surfaces stale.
22−- When changing user-visible behavior, keep `README.md` and `README.zh-CN.md` consistent.
23−- `AGENTS.md` and `CLAUDE.md` are active project-level instruction surfaces. Keep them aligned; use `CLAUDE.md` only for Claude Code-specific notes and point back here for shared rules.
24−- New features or behavior changes must add or update relevant tests when an automated test surface exists. If no automated suite exists, run the closest static/syntax checks and document why no test was added.
25−- Record stable cross-session decisions in the native project memory surface when one is available. If none is available, record only agent-facing rules in `AGENTS.md`/`CLAUDE.md`; do not invent a repo-local memory file without an explicit workflow decision.
26−
27−## Validation
28−
29−Run the standing consistency guard first; it subsumes the Python byte-compile check. Then use the closest checks for the changed surface:
30−
31−- `bash scripts/validate.sh`
32−- `bash -n scripts/install-codex.sh scripts/install-cursor.sh scripts/install-opencode.sh scripts/install-agents.sh scripts/install-all.sh`
33−- Targeted `rg` checks for newly required workflow language
34−- `git diff --check`
7+- The optional Claude Code sub-agent prompts live in `plugins/spec-driven-develop/agents/`, including `code-reviewer.md`, the per-lane review-and-fix sub-agent used in the execution loop.
8+- The plugin is skills-only: workflows are invoked through `plugins/spec-driven-develop/skills/`, not slash commands.
9+- When the shared workflow rules change, update Claude Code-specific prompts only where their execution contract also changes.
10+- Do not duplicate shared policy here. Put durable cross-agent rules in `AGENTS.md`; use Claude Code's native project memory surface for stable project facts when available.
3511
