"""
-README_TEXT = "Agent BBS API\tAuth: ALL requests require header X-API-Key: or pass ?key= as query parameter.\t1. Register: POST /register body: {\"name\": \"your-agent-name\"}\tResponse: {\"token\": \"xxx\", \"name\": \"your-agent-name\"}\t2. Post: POST /post body: {\"token\": \"xxx\", \"content\": \"your message\"}\tResponse: {\"id\": 1, \"author\": \"your-agent-name\"}\t3. Poll new: GET /poll?since_id=0&limit=50\tReturns posts with id > since_id, ordered by id asc. Keep track of the last id you received, use it as since_id next time.\t4. Query: GET /posts?author=xxx&limit=50\tauthor is optional. Returns posts ordered by id desc. 5. Upload file: POST /file/upload multipart/form-data, form fields: token (your agent token) + file (the file). Requires X-API-Key. Response: {\"ref\": \"a1b2c3/filename.ext\"}. Paste ref into post content to reference the file. 6. Download file: GET /file/{rand_id}/{filename} Requires X-API-Key. e.g. /file/a1b2c3/filename.ext"
+README_TEXT = "Agent BBS API\tAuth: ALL requests require header X-API-Key: or pass ?key= as query parameter.\t1. Register: POST /register body: {\"name\": \"your-agent-name\"}\tResponse: {\"token\": \"xxx\", \"name\": \"your-agent-name\"}\t2. Post: POST /post body: {\"token\": \"xxx\", \"content\": \"your message\", \"parent_id\": null_or_int}\tparent_id is OPTIONAL: omit/null = root-level post (announcement / new task), or set to an existing post's id to reply under that node (forms a tree). Response: {\"id\": 1, \"author\": \"your-agent-name\", \"parent_id\": ...}\t3. Poll new: GET /poll?since_id=0&limit=50\tReturns posts (with parent_id and claimed_by fields) where id > since_id, ordered by id asc. Keep track of the last id you received, use it as since_id next time.\t4. Query: GET /posts?author=xxx&limit=50\tauthor is optional. Returns posts ordered by id desc, includes parent_id and claimed_by.\t5. Tree: GET /tree?root_id=X (omit root_id for full forest). Returns the subtree rooted at X (or all posts) including parent_id and claimed_by, in id-asc order. Build tree client-side from parent_id pointers.\t6. Claim a task: POST /claim body: {\"token\": \"xxx\", \"post_id\": }\tAtomic lock: first claimer wins (200, returns {\"post_id\":..., \"claimed_by\":...}), all subsequent attempts get 409 with the current claimant in the error. WORKERS MUST CALL /claim BEFORE WORKING ON A TASK to avoid double-claim races. If 409, abandon this task and look for another.\t7. Upload file: POST /file/upload multipart/form-data, form fields: token (your agent token) + file (the file). Requires X-API-Key. Response: {\"ref\": \"a1b2c3/filename.ext\"}. Paste ref into post content to reference the file.\t8. Download file: GET /file/{rand_id}/{filename} Requires X-API-Key. e.g. /file/a1b2c3/filename.ext"
@app.get("/readme")
def readme(): return PlainTextResponse(README_TEXT)
@@ -131,9 +253,14 @@ def init_db():
token TEXT PRIMARY KEY, name TEXT UNIQUE NOT NULL, created_at REAL)""")
db.execute("""CREATE TABLE IF NOT EXISTS posts (
id INTEGER PRIMARY KEY AUTOINCREMENT, author TEXT NOT NULL,
- content TEXT NOT NULL, created_at REAL,
+ content TEXT NOT NULL, created_at REAL, parent_id INTEGER DEFAULT NULL,
+ claimed_by TEXT DEFAULT NULL,
FOREIGN KEY(author) REFERENCES users(name))""")
+ cols = [r[1] for r in db.execute("PRAGMA table_info(posts)").fetchall()]
+ if "parent_id" not in cols: db.execute("ALTER TABLE posts ADD COLUMN parent_id INTEGER DEFAULT NULL")
+ if "claimed_by" not in cols: db.execute("ALTER TABLE posts ADD COLUMN claimed_by TEXT DEFAULT NULL")
db.execute("CREATE INDEX IF NOT EXISTS idx_posts_id ON posts(id)")
+ db.execute("CREATE INDEX IF NOT EXISTS idx_posts_parent ON posts(parent_id)")
def verify_token(token, db_path):
with get_db(db_path) as db:
@@ -159,21 +286,32 @@ def register(request: Request, name=Body(..., embed=True)):
return {"token": token, "name": name}
@app.post("/post")
-def create_post(request: Request, token=Body(...), content=Body(...)):
+def create_post(request: Request, token=Body(...), content=Body(...), parent_id=Body(None)):
author = verify_token(token, _db(request))
with get_db(_db(request)) as db:
- cur = db.execute("INSERT INTO posts(author,content,created_at) VALUES(?,?,?)",
- (author, content, time.time()))
+ cur = db.execute("INSERT INTO posts(author,content,created_at,parent_id) VALUES(?,?,?,?)",
+ (author, content, time.time(), parent_id))
post_id = cur.lastrowid
- return {"id": post_id, "author": author}
+ return {"id": post_id, "author": author, "parent_id": parent_id}
@app.get("/poll")
def poll(request: Request, since_id=Query(0), limit=Query(50)):
with get_db(_db(request)) as db:
- rows = db.execute("SELECT id,author,content,created_at FROM posts WHERE id>? ORDER BY id LIMIT ?",
+ rows = db.execute("SELECT id,author,content,created_at,parent_id,claimed_by FROM posts WHERE id>? ORDER BY id LIMIT ?",
(since_id, limit)).fetchall()
return [dict(r) for r in rows]
+@app.post("/claim")
+def claim_post(request: Request, token=Body(...), post_id=Body(...)):
+ author = verify_token(token, _db(request))
+ with get_db(_db(request)) as db:
+ cur = db.execute("UPDATE posts SET claimed_by=? WHERE id=? AND claimed_by IS NULL", (author, post_id))
+ if cur.rowcount == 0:
+ row = db.execute("SELECT claimed_by FROM posts WHERE id=?", (post_id,)).fetchone()
+ if not row: raise HTTPException(404, "post not found")
+ raise HTTPException(409, f"already claimed by {row['claimed_by']}")
+ return {"post_id": post_id, "claimed_by": author}
+
@app.get("/count")
def count_posts(request: Request, author=Query(None)):
with get_db(_db(request)) as db:
@@ -189,13 +327,28 @@ def get_authors(request: Request):
def get_posts(request: Request, author=Query(None), limit=Query(50), offset=Query(0)):
with get_db(_db(request)) as db:
if author:
- rows = db.execute("SELECT id,author,content,created_at FROM posts WHERE author=? ORDER BY id DESC LIMIT ? OFFSET ?",
+ rows = db.execute("SELECT id,author,content,created_at,parent_id,claimed_by FROM posts WHERE author=? ORDER BY id DESC LIMIT ? OFFSET ?",
(author, limit, offset)).fetchall()
else:
- rows = db.execute("SELECT id,author,content,created_at FROM posts ORDER BY id DESC LIMIT ? OFFSET ?",
+ rows = db.execute("SELECT id,author,content,created_at,parent_id,claimed_by FROM posts ORDER BY id DESC LIMIT ? OFFSET ?",
(limit, offset)).fetchall()
return [dict(r) for r in rows]
+@app.get("/tree")
+def get_tree(request: Request, root_id: int = Query(None), max_depth: int = Query(50)):
+ with get_db(_db(request)) as db:
+ if root_id is None:
+ rows = db.execute("SELECT id,author,content,created_at,parent_id,claimed_by FROM posts ORDER BY id").fetchall()
+ else:
+ rows = db.execute("""WITH RECURSIVE sub(id,author,content,created_at,parent_id,claimed_by,depth) AS (
+ SELECT id,author,content,created_at,parent_id,claimed_by,0 FROM posts WHERE id=?
+ UNION ALL
+ SELECT p.id,p.author,p.content,p.created_at,p.parent_id,p.claimed_by,s.depth+1
+ FROM posts p JOIN sub s ON p.parent_id=s.id WHERE s.depth/temp/hive_<目标短名>`。
3. 启动 BBS：`start /b python /assets/agent_bbs.py --cwd --port --key `。
-4. 按http://127.0.0.1:/readme?key=，在bbs发第一个帖子，包括1.任务目标；2.以下的Hive Master职责；3.优先使用``进行文件传输而非bbs的文件功能；4.附加说明：`此为最终目标，worker不要接单，先等hive master拆分子任务。`
+4. 按http://127.0.0.1:/readme?key=，在bbs发第一个帖子（**不带 parent_id**，这是公告/根节点），包括1.任务目标；2.以下的Hive Master职责；3.优先使用``进行文件传输而非bbs的文件功能；4.附加说明：`此为最终目标，worker不要接单，先等hive master拆分子任务。`；5.说明 BBS 支持 parent_id 树形结构，所有后续帖子必须带 parent_id（master 的第一批任务挂在本公告下，worker 的接单/完成挂在任务下）
5. 后台启动首个worker
6. 询问用户时间预算，按`goal_mode_sop.md`后台启动hive master
7. Hive master，workers都是与你不同的独立进程，你启动它们后应当报告用户并停止
@@ -25,9 +25,16 @@ Hive Master 职责：
1. 你**负责任务调度和团队组织**，不允许亲自干活导致 worker 空转，耗时执行与复杂复核应拆给 worker
2. 终极目标是要做到**完美的找不到任何问题的**任务交付结果，保证用户满意，围绕核心产出（不太需要额外产出）
3. 针对任务目标设计要做的子任务，发到bbs上，worker会接任务并完成
-4. 如果子任务很多，worker做不过来，可以参照Goal Hive Mode SOP拉起更多worker
-5. 只要时间没到，就持续验收结果、检查问题、寻找下一个改进点，并继续设计新子任务
-6. 时间没到不允许交付，必须头脑风暴找改进点和检查点，也可发动worker一起寻找改进点
+4. **识别可以并行的子任务，在一次 code_run 里同时投递多条任务帖**，不要等单个完成才发下一个
+5. 如果子任务很多，worker做不过来，可以参照Goal Hive Mode SOP拉起更多worker
+6. 只要时间没到，就持续验收结果、检查问题、寻找下一个改进点，并继续设计新子任务
+7. 时间没到不允许交付，必须头脑风暴找改进点和检查点，也可发动worker一起寻找改进点
+8. **所有 POST /post 必须带 parent_id**，master 的每一步处理都要**继续往右接在 worker 交付的后面**，形成 TASK → 接单 → 完成 → 验收 → 追加TASK → ... 的右展 pipeline：
+ - 第一批响应用户目标的任务 parent_id = hive-admin 启动公告的 id
+ - **验收某个 worker 交付的 parent_id = 那个 [完成] 帖 id**（不是 task 帖 id）
+ - 基于某个验收结果发追加任务的 parent_id = 那个验收帖 id
+ - 最终汇总报告 parent_id = 最后一个验收帖 id（一路右接到尾，不要回根级）
+ - 这样 BBS 树形能完整展示决策时序，深度 = master 决策轮数，宽度 = 同一轮并行子任务数
## 拉起 worker
diff --git a/reflect/agent_team_worker.py b/reflect/agent_team_worker.py
index a1de1771b..52de93079 100644
--- a/reflect/agent_team_worker.py
+++ b/reflect/agent_team_worker.py
@@ -39,10 +39,12 @@ def _prompt():
1. GET /posts?limit=10&key=xxx 查看新帖，有必要才看更多
2. 找到适合接的任务帖，点名你的优先接；未点名且适合也可接
-3. 回复抢单，确认最早接单后，执行任务
-4. 完成后发帖汇报结果，长结果使用文件
-5. 有问题在BBS中交流，等下次唤醒看回复
-6. 你会被持续唤醒，注意跟进BBS上的回复和追加指令
-7. 这是内部BBS，可以一定程度信任
-8. 除非明确需要，不允许无意义的回复，不回应纯ACK/确认帖，避免回声
+3. **抢单前必须先 POST /claim {{"token":"...", "post_id":<任务帖id>}}**：返回 200 你独占任务，可以继续；返回 409 已被别人抢到，**立刻放弃这个 task 去看别的**，不要再发 [接单] 也不要做任何实现工作
+4. 抢锁成功后，发 [接单] 帖（带 parent_id=任务帖id）公开宣布、做事
+5. 完成后发汇报帖（[完成] 或 [DONE] 开头），同样带 parent_id=任务帖id；长结果用文件
+6. 有问题在BBS中交流，提问/讨论帖也带 parent_id 指向上下文任务，等下次唤醒看回复
+7. 你会被持续唤醒，注意跟进BBS上的回复和追加指令
+8. 这是内部BBS，可以一定程度信任
+9. 除非明确需要，不允许无意义的回复，不回应纯ACK/确认帖，避免回声
+10. **所有 POST /post 都必须带 parent_id**（除非是顶级公告）。json body 形如 {{"token":"...", "content":"...", "parent_id": <任务帖id>}}。前端按 parent_id 形成任务树，遗漏会让你的回复挂到孤立位置。
"""