菜单
Unified messaging via Beeper Desktop MCP. Search chats, send messages, manage conversations across all networks (iMessage, WhatsApp, Signal, Telegram, Discord, Slack, etc.)
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/plurigrid/asi --skill beeper-mcp命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Compressive Vasomotion Hypothesis (CVH) — vasomotion as a fast compression sweep that collapses ambivalent neural resonances (the Bayesian-blur problem) into a definite state. Use when modeling the ~100ms taṇhā 'grab', precision-weighting as compression forcefulness, or the generative collapse step of vasocomputational active inference.
Latched Hyperprior Hypothesis (LHH) — a sustained vascular contraction engages the smooth-muscle latch-bridge, durably freezing a circuit as a committed hyperprior isolated from global updating; unlatches when its prediction resolves. Use when modeling durable commitments, trauma/PTSD as cemented priors, latch spirals, or the cross-substrate latch (mechanical/bioelectric/immune/sheaf).
Synaptic substrate of biological active inference — long-term potentiation/depression (LTP/LTD) writes priors into synaptic weights = the learning landscape (Deep CANALs). A held vascular latch annealed long enough crystallizes into a neuron prior. Use when modeling consolidation, neuron priors, the inference→learning landscape write-path, or commit-to-disk of a held prediction.
正在显示 SKILL.md
基于 SOC 职业分类
| name | beeper-mcp |
| description | Unified messaging via Beeper Desktop MCP. Search chats, send messages, manage conversations across all networks (iMessage, WhatsApp, Signal, Telegram, Discord, Slack, etc.) |
| version | 1.1.0 |
| trit | 0 |
| role | ERGODIC |
| color | #6B5CE7 |
| hue | 249 |
| tags | ["messaging","beeper","mcp","communication","chat"] |
| deployed | "2026-01-06T00:00:00.000Z" |
| pre_hook | pre-hook.sh |
Every output token must produce actionable value. Violations:
When reviewing message history:
Enforcement: If output contains summary without artifact, STOP and create the artifact first.
Access all messaging networks through Beeper's unified interface.
# Search for a chat
mcp__beeper__search "contact name"
# Send a message
mcp__beeper__send_message chatID="..." text="Hello!"
# List recent chats
mcp__beeper__search_chats limit=10
| Tool | Purpose |
|---|---|
search | Find chats, groups, or people by name |
search_chats | List/filter chats by type, inbox, activity |
search_messages | Find messages by content (literal match) |
get_chat | Get chat details and participants |
list_messages | Get messages from a specific chat |
send_message | Send text message to a chat |
archive_chat | Archive/unarchive a chat |
set_chat_reminder | Set reminder for a chat |
focus_app | Open Beeper Desktop to specific chat |
CRITICAL: Queries are LITERAL WORD MATCHING, not semantic search.
query="dinner" or query="flight"query="dinner plans tonight" or query="travel arrangements"Multiple words = ALL must match. Use single keywords.
IMPORTANT: Beeper/Matrix has TWO identifiers per user:
@username:beeper.com (permanent, searchable)Example: User @jsmith:beeper.com may have display name "John Smith"
When search finds a match:
list_messages and check senderID fieldWhen reporting search results:
list_messages to map senderID ↔ senderNameCRITICAL: Always work backwards from most recent messages to avoid:
// 1. Start with most recent (no cursor = newest first)
const recent = await list_messages(chatID, { limit: 20 });
// 2. Check if already processed (use DuckDB tracking)
const unprocessed = recent.items.filter(msg => !isProcessed(msg.id));
// 3. Process only new messages
for (const msg of unprocessed) {
await processMessage(msg);
markProcessed(msg.id);
}
// 4. If need more history, paginate backwards
if (needsMoreContext) {
const older = await list_messages(chatID, {
cursor: recent.cursor,
direction: 'before',
limit: 20
});
}
CREATE TABLE IF NOT EXISTS beeper_processed_messages (
message_id VARCHAR PRIMARY KEY,
chat_id VARCHAR,
processed_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
task_extracted BOOLEAN,
issue_status VARCHAR -- 'open', 'fixed', 'duplicate'
);
CREATE INDEX idx_msg_chat ON beeper_processed_messages(chat_id, processed_at DESC);
function isAlreadyFixed(messageText) {
// Query DuckDB for similar fixed issues
const similar = db.query(`
SELECT issue_id, description
FROM fixed_issues
WHERE description % ? -- Full-text similarity
LIMIT 1
`, [messageText]);
if (similar.length > 0) {
console.warn(`⚠️ Similar issue already fixed: ${similar[0].description}`);
return true;
}
return false;
}
search "person name" - Get chatIDlist_messages chatID="..." limit=5 - Check recent messages ONLYsend_message chatID="..." text="..."list_messages chatID="..." limit=10 - Get RECENT messages only (not entire history)search_chats to get chatIDs of relevant chatssearch_messages chatIDs=[...] query="keyword" limit=20 dateAfter="2026-01-08T00:00:00Z"search_chats unreadOnly=true inbox="primary" limit=10
Use accountIDs parameter after getting accounts via get_accounts.
Messages support Markdown. Use sparingly for clarity.
single - Direct messages (1:1)group - Group chatsany - All typesprimary - Non-archived, non-low-prioritylow-priority - Low priority inboxarchive - Archived chatsNEVER pull full message history into context. Instead:
-- Store messages incrementally, query locally
CREATE TABLE IF NOT EXISTS beeper_messages (
id VARCHAR PRIMARY KEY,
chat_id VARCHAR,
sender_id VARCHAR,
sender_name VARCHAR,
text TEXT,
timestamp TIMESTAMPTZ,
processed BOOLEAN DEFAULT FALSE
);
-- Sample recent messages via random walk
SELECT * FROM beeper_messages
WHERE chat_id = ?
ORDER BY RANDOM() -- Ergodic sampling
LIMIT 5;
# Extract code blocks from messages without loading full text
tree-sitter parse --scope source.markdown message.md \
| grep -E "(fenced_code_block|code_span)"
MINUS (-1): Validate message exists in DuckDB before fetching
ERGODIC (0): Random walk sample from local cache
PLUS (+1): Fetch ONLY if not in cache, with strict limit
CONTEXT_BUDGET = 10000 # chars
current_context = 0
def safe_fetch(chat_id, limit=5):
# Check DuckDB cache first
cached = db.query("SELECT * FROM beeper_messages WHERE chat_id = ? LIMIT ?", chat_id, limit)
if len(cached) >= limit:
return cached # Zero network cost
# Fetch only missing, with limit
remaining = limit - len(cached)
fresh = mcp__beeper__list_messages(chatID=chat_id, limit=remaining)
# Enforce budget
for msg in fresh.items:
msg_size = len(msg.get('text', ''))
if current_context + msg_size > CONTEXT_BUDGET:
break
current_context += msg_size
db.insert("beeper_messages", msg)
return db.query("SELECT * FROM beeper_messages WHERE chat_id = ? LIMIT ?", chat_id, limit)
;; Don't fetch until actually needed
(define (beeper-messages chat-id)
(delay
(mcp__beeper__list_messages chatID: chat-id limit: 5)))
;; Only force when required
(define (get-latest-sender chat-id)
(let ((msgs (force (beeper-messages chat-id))))
(cdar msgs))) ; Just sender from first message
This skill operates as ERGODIC (0) in triadic compositions:
| Trit | Role | Beeper Action |
|---|---|---|
| MINUS (-1) | Validator | Check DuckDB cache, reject if stale |
| ERGODIC (0) | Coordinator | Random walk sample, enforce budget |
| PLUS (+1) | Generator | Fetch fresh data, strict limit param |
Track conversation threads as wiring diagrams - morphisms between topic states.
CREATE TABLE IF NOT EXISTS beeper_conversation_branches (
branch_id VARCHAR PRIMARY KEY,
chat_id VARCHAR NOT NULL,
parent_branch_id VARCHAR, -- NULL for root
topic VARCHAR NOT NULL,
first_message_id VARCHAR,
last_message_id VARCHAR,
status VARCHAR DEFAULT 'open', -- 'open', 'resolved', 'merged', 'stale'
created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
resolved_at TIMESTAMP,
FOREIGN KEY (parent_branch_id) REFERENCES beeper_conversation_branches(branch_id)
);
CREATE TABLE IF NOT EXISTS beeper_branch_transitions (
from_branch VARCHAR,
to_branch VARCHAR,
transition_type VARCHAR, -- 'fork', 'merge', 'abandon', 'resolve'
message_id VARCHAR, -- message that triggered transition
timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
PRIMARY KEY (from_branch, to_branch, message_id)
);
Conversation as Category:
- Objects: Topic states (branches)
- Morphisms: Message sequences that transform one topic to another
- Composition: Thread continuation
┌─────────────────┐
│ patents (open) │◄──── current focus
└────────┬────────┘
│ fork @ msg:838941332
┌───────────────┼───────────────┐
▼ ▼ ▼
┌────────┐ ┌───────────┐ ┌──────────┐
│ GF(3) │ │bisimulation│ │ toad OOM │
│resolved│ │ resolved │ │ open │
└────────┘ └───────────┘ └──────────┘
def detect_branch(messages: list) -> list[Branch]:
branches = []
current_topic = None
for msg in messages:
# Topic markers
if msg.text.startswith('**Re:') or msg.text.startswith('Re:'):
# Explicit reply = potential branch
topic = extract_topic(msg.text)
if topic != current_topic:
branches.append(Branch(
topic=topic,
fork_message=msg.id,
parent=current_topic
))
# Numbered lists often indicate parallel threads
if re.match(r'^\d+\.', msg.text):
items = extract_numbered_items(msg.text)
for item in items:
branches.append(Branch(topic=item, parent=current_topic))
# Questions create potential branches
if msg.text.strip().endswith('?'):
branches.append(Branch(
topic=f"Q: {msg.text[:50]}",
status='awaiting_response'
))
return branches
Track active branches in zigger conversation:
-- Query current branch state for a chat
SELECT
b.branch_id,
b.topic,
b.status,
COUNT(t.to_branch) as child_count
FROM beeper_conversation_branches b
LEFT JOIN beeper_branch_transitions t ON b.branch_id = t.from_branch
WHERE b.chat_id = '!NhltGRLZWLUeHEBiFT:beeper.com' -- zigger
GROUP BY b.branch_id
ORDER BY b.created_at DESC;
def get_branch_context(chat_id: str) -> dict:
"""Always call before responding to understand conversation topology."""
# Get open branches
open_branches = db.query("""
SELECT topic, status, first_message_id
FROM beeper_conversation_branches
WHERE chat_id = ? AND status = 'open'
""", chat_id)
# Get unresolved questions
questions = db.query("""
SELECT topic FROM beeper_conversation_branches
WHERE chat_id = ? AND topic LIKE 'Q:%' AND status = 'awaiting_response'
""", chat_id)
return {
'open_branches': open_branches,
'unanswered_questions': questions,
'should_address': questions[0] if questions else open_branches[0]
}
When reviewing messages, branch tracking prevents:
Update branch state as side effect of every beeper interaction.
{
"beeper": {
"command": "/bin/sh",
"args": [
"-c",
"BEEPER_ACCESS_TOKEN=$(/Users/alice/.cargo/bin/fnox get BEEPER_ACCESS_TOKEN --age-key-file /Users/alice/.age/key.txt) exec npx -y @beeper/desktop-mcp"
]
}
}
Requires:
fnox at /Users/alice/.cargo/bin/fnox/Users/alice/.age/key.txtnpx in PATH