@synapse-research/synapse 0.2.11 → 0.2.13

package/dist/public/skill/aris-synapse-sync/SKILL.md

@@ -1,368 +0,0 @@
-# ARIS → Synapse Dashboard Sync
-
-## Purpose
-
-When running ARIS research workflows (idea-discovery, experiment-bridge, run-experiment, auto-review-loop, paper-writing, research-pipeline), **automatically sync progress to Synapse** for real-time dashboard visualization.
-
-This skill is a **sidecar** — it adds observability without changing ARIS behavior. ARIS controls the research flow; Synapse provides the dashboard.
-
-## Golden Rule
-
-**Sync failures MUST NOT block ARIS workflows.** If any Synapse MCP call fails, log it and continue. ARIS execution always takes priority.
-
-## Prerequisites
-
-- Synapse MCP server configured in `.mcp.json` or `~/.claude/settings.json`
-- Agent API key with roles: `pre_research`, `research`, `experiment`, `report`
-- (Optional) A pre-created Synapse ResearchProject UUID
-
-## State File
-
-Maintain `.aris-synapse-sync.json` in the working directory. This maps ARIS artifacts to Synapse UUIDs.
-
-```json
-{
-  "projectUuid": null,
-  "mainExperimentUuid": null,
-  "researchQuestions": {},
-  "documents": {},
-  "reviewRound": 0,
-  "lastSyncAt": null
-}
-```
-
-**On first sync:** create this file. **On subsequent syncs:** read, update, write back.
-
----
-
-## Checkpoint Protocol
-
-Execute these checkpoints at the corresponding ARIS workflow stages. Each checkpoint is independent — execute whichever checkpoints apply to the current workflow.
-
-### Checkpoint 0: Session Init
-
-**When:** At the START of any ARIS workflow (before any ARIS skill execution).
-
-```
-1. Read .aris-synapse-sync.json (or create if missing)
-2. Call synapse_checkin()
-3. If projectUuid is null:
-   a. If user provided a Synapse project UUID → use it
-   b. Otherwise → synapse_create_research_project({
-        name: "ARIS: <research direction/topic>",
-        description: "<1-2 sentence research goal>"
-      })
-   c. Save projectUuid to state file
-4. Log: "Synapse sync initialized → project <uuid>"
-```
-
-### Checkpoint 1: Literature Sync
-
-**When:** After `/research-lit` or any literature search completes.
-
-For each paper discovered:
-```
-synapse_add_related_work({
-  researchProjectUuid: <projectUuid>,
-  title: <paper title>,
-  url: <arxiv URL or DOI link>,
-  authors: <author list as string>,
-  abstract: <abstract text>,
-  arxivId: <arxiv ID if available>,
-  year: <publication year>,
-  source: "arxiv"  // or "semantic_scholar", "deepxiv"
-})
-```
-
-If a deep research / literature review report is produced:
-```
-synapse_save_deep_research_report({
-  researchProjectUuid: <projectUuid>,
-  title: "ARIS Literature Review: <topic>",
-  content: <full review content as markdown>
-})
-```
-
-**Note:** `synapse_add_related_work` has built-in dedup (returns `isNew: false` for duplicates). Safe to re-sync.
-
-### Checkpoint 2: Ideas Sync
-
-**When:** After `/idea-creator` or `/idea-discovery` produces ranked ideas (IDEA_REPORT.md).
-
-For each generated idea:
-```
-synapse_create_research_question({
-  researchProjectUuid: <projectUuid>,
-  title: "Idea #<rank>: <idea title>",
-  content: "<idea summary>\n\n**Confidence:** <score>\n**Novelty:** <assessment>\n**Feasibility:** <assessment>"
-})
-→ Save { "<idea-slug>": "<researchQuestionUuid>" } to state file
-```
-
-### Checkpoint 3: Idea Selected → Experiment Created
-
-**When:** After an idea is selected for execution (user selects or auto-select #1).
-
-```
-synapse_propose_experiment({
-  researchProjectUuid: <projectUuid>,
-  title: <selected idea title>,
-  description: <idea description + hypothesis + expected outcome>,
-  researchQuestionUuid: <mapped from state file if available>,
-  priority: "high"
-})
-→ Save mainExperimentUuid to state file
-```
-
-**Note:** If the Synapse project has autonomous loop enabled, this creates the experiment in `pending_start` and auto-assigns it to the agent. Otherwise it lands in `pending_review` (visible on the board either way).
-
-### Checkpoint 4: Experiment Plan Sync
-
-**When:** After `/experiment-plan` or `/experiment-bridge` produces EXPERIMENT_PLAN.md.
-
-```
-synapse_create_document({
-  researchProjectUuid: <projectUuid>,
-  type: "methodology",
-  title: "ARIS Experiment Plan: <topic>",
-  content: <full experiment plan as markdown>
-})
-→ Save { "EXPERIMENT_PLAN": "<documentUuid>" } to state file
-```
-
-### Checkpoint 5: Experiment Execution
-
-**When:** During `/run-experiment` or `/experiment-bridge` execution phases.
-
-**5a. Experiment starts:**
-```
-synapse_start_experiment({
-  experimentUuid: <mainExperimentUuid>,
-  workingNotes: "ARIS experiment execution started.\nGPU: <gpu info>\nEstimated time: <budget>"
-})
-```
-
-If `synapse_start_experiment` fails (e.g. wrong status), fall back to progress reporting only.
-
-**5b. During execution (periodic, every significant milestone):**
-```
-synapse_report_experiment_progress({
-  experimentUuid: <mainExperimentUuid>,
-  message: <current status — e.g. "Sanity check passed ✓" or "Epoch 5/20 | loss=0.342 | acc=0.76">,
-  phase: <"sanity_check" | "baseline" | "main_experiment" | "ablation">,
-  liveStatus: "running"
-})
-```
-
-Report at these ARIS milestones:
-- Sanity check start/pass/fail
-- Each experiment run start
-- Training progress (every ~25% or significant metric change)
-- Run completion with key metrics
-- Auto-debug retries
-
-**5c. Baseline registration (if applicable):**
-```
-synapse_create_baseline({
-  researchProjectUuid: <projectUuid>,
-  name: "ARIS Baseline: <method name>",
-  metrics: { "accuracy": 0.82, "f1": 0.79, ... },
-  experimentUuid: <mainExperimentUuid>
-})
-```
-
-### Checkpoint 6: Review Loop Sync
-
-**When:** During `/auto-review-loop`, after each review round completes.
-
-**6a. Review received (after Phase A-B):**
-```
-synapse_add_comment({
-  targetType: "experiment",
-  targetUuid: <mainExperimentUuid>,
-  content: "## 📋 Review Round <N>\n\n**Score:** <X>/10\n**Verdict:** <ready|almost|not ready>\n\n### Weaknesses\n<ranked list>\n\n### Required Fixes\n<action items>\n\n---\n*Reviewer: <codex|oracle-pro>*"
-})
-
-synapse_report_experiment_progress({
-  experimentUuid: <mainExperimentUuid>,
-  message: "Review Round <N>: <score>/10 — <verdict>",
-  phase: "review_round_<N>",
-  liveStatus: "running"
-})
-
-→ Update reviewRound in state file
-```
-
-**6b. Fixes implemented (after Phase C-D):**
-```
-synapse_report_experiment_progress({
-  experimentUuid: <mainExperimentUuid>,
-  message: "Round <N> fixes applied: <summary of changes and new experiments run>",
-  phase: "review_fixes_<N>"
-})
-```
-
-**6c. Debate transcript (hard/nightmare mode, after Phase B.5-B.6):**
-```
-synapse_add_comment({
-  targetType: "experiment",
-  targetUuid: <mainExperimentUuid>,
-  content: "## ⚖️ Debate Round <N>\n\n<debate transcript with SUSTAINED/OVERRULED verdicts>"
-})
-```
-
-### Checkpoint 7: Experiment Complete
-
-**When:** After `/auto-review-loop` finishes (all rounds done or score threshold met), or after `/run-experiment` completes without review loop.
-
-```
-synapse_submit_experiment_results({
-  experimentUuid: <mainExperimentUuid>,
-  outcome: <score >= 6 ? "positive" : "negative">,
-  experimentResults: {
-    "finalScore": <last review score>,
-    "totalRounds": <N>,
-    "scoreProgression": [5.0, 6.5, 6.8, 7.5],
-    "verdict": <final verdict>,
-    "keyMetrics": { <metric: value pairs> },
-    "summary": <1-2 paragraph result summary>
-  }
-})
-```
-
-### Checkpoint 8: Document Sync
-
-**When:** After `/paper-writing`, `/paper-compile`, or report generation.
-
-**8a. Narrative report:**
-```
-doc = state.documents["NARRATIVE_REPORT"]
-if doc exists:
-  synapse_update_document({ documentUuid: doc, content: <updated content> })
-else:
-  synapse_create_document({
-    researchProjectUuid: <projectUuid>,
-    type: "results_report",
-    title: "ARIS Narrative Report: <topic>",
-    content: <NARRATIVE_REPORT.md content>
-  })
-  → Save to state file
-```
-
-**8b. Paper draft:**
-```
-synapse_create_document({
-  researchProjectUuid: <projectUuid>,
-  type: "other",
-  title: "ARIS Paper Draft: <paper title>",
-  content: <paper content or "Paper compiled. See local PDF: <path>">
-})
-```
-
-**8c. Auto-improvement round results:**
-```
-synapse_update_document({
-  documentUuid: <paper doc uuid>,
-  content: <updated paper content after improvement round>
-})
-// Document versioning auto-increments (v1 → v2 → v3)
-```
-
----
-
-## Workflow-to-Checkpoint Mapping
-
-Quick reference for which checkpoints apply to each ARIS workflow:
-
-| ARIS Workflow | Checkpoints |
-|---------------|-------------|
-| `/idea-discovery` | 0 → 1 → 2 |
-| `/research-lit` | 0 → 1 |
-| `/idea-creator` | 0 → 2 |
-| `/experiment-plan` | 0 → 4 |
-| `/experiment-bridge` | 0 → 3 → 4 → 5 |
-| `/run-experiment` | 0 → 5 → 7 |
-| `/auto-review-loop` | 0 → 6 → 7 |
-| `/paper-writing` | 0 → 8 |
-| `/paper-compile` | 0 → 8 |
-| `/auto-paper-improvement-loop` | 0 → 8 |
-| `/research-pipeline` | 0 → 1 → 2 → 3 → 4 → 5 → 6 → 7 → 8 (full chain) |
-| `/result-to-claim` | 0 → 7 (update experiment results with claims) |
-
-## Error Handling
-
-```
-For every Synapse MCP call:
-  try:
-    result = call synapse tool
-    update state file
-  catch:
-    log: "⚠️ Synapse sync failed at checkpoint <N>: <error>"
-    continue ARIS workflow (NEVER block)
-```
-
-Common failures and responses:
-- **MCP server unreachable**: Log warning, disable sync for this session (set `syncEnabled: false` in state)
-- **Auth error (401/403)**: Log "API key missing or invalid", disable sync
-- **Project not found**: Re-run checkpoint 0 to create project
-- **Experiment status conflict**: Fall back to progress reporting and comments only (skip `start_experiment` / `submit_results`)
-- **Rate limit**: Add 2s delay between batch calls (e.g. multiple `add_related_work`)
-
-## State File Recovery
-
-If `.aris-synapse-sync.json` is lost or corrupted:
-
-1. Call `synapse_checkin()` to get agent identity
-2. Call `synapse_list_research_projects()` to find existing ARIS project (match by name prefix "ARIS:")
-3. Call `synapse_get_assigned_experiments()` to find active experiment
-4. Reconstruct state file from Synapse data
-5. Continue sync from current ARIS workflow state
-
-## Setup Instructions
-
-### 1. Configure Synapse MCP
-
-Add to `.mcp.json` in your research project directory:
-```json
-{
-  "mcpServers": {
-    "synapse": {
-      "type": "streamable-http",
-      "url": "https://<synapse-host>/api/mcp",
-      "headers": {
-        "Authorization": "Bearer syn_<your-api-key>"
-      }
-    }
-  }
-}
-```
-
-### 2. Create Synapse Agent
-
-In Synapse web UI (`/agents`):
-- Create agent with roles: `pre_research`, `research`, `experiment`, `report`
-- Type: `claude_code`
-- Generate API key (prefix: `syn_`)
-
-### 3. (Optional) Pre-create Project
-
-Create a ResearchProject in Synapse web UI. Copy its UUID and pass when starting ARIS:
-```
-> /research-pipeline "your topic" — synapse-project: <uuid>
-```
-
-Or let the sync skill auto-create the project at checkpoint 0.
-
-### 4. Install This Skill
-
-```bash
-# Copy to Claude Code skills directory
-cp -r aris-synapse-sync/ ~/.claude/skills/aris-synapse-sync/
-```
-
-### 5. Run ARIS Normally
-
-```
-> /idea-discovery "efficient attention for long sequences"
-# Sync happens automatically at each checkpoint
-```

package/dist/public/skill/package.json

@@ -1,79 +0,0 @@
-{
-  "name": "synapse-skill",
-  "version": "0.1.0",
-  "description": "Synapse AI Agent collaboration platform skill. Supports PM, Developer, and Admin roles via MCP tools for the Idea-Proposal-Task workflow.",
-  "author": "synapse",
-  "license": "AGPL-3.0",
-  "keywords": [
-    "synapse",
-    "skill",
-    "project-management",
-    "ai-agent",
-    "mcp",
-    "pm",
-    "developer",
-    "admin",
-    "collaboration"
-  ],
-  "synapse": {
-    "category": "project-management",
-    "mcp_server": "synapse",
-    "files": {
-      "SKILL.md": "/skill/SKILL.md",
-      "references/00-common-tools.md": "/skill/references/00-common-tools.md",
-      "references/01-setup.md": "/skill/references/01-setup.md",
-      "references/02-pm-workflow.md": "/skill/references/02-pm-workflow.md",
-      "references/03-developer-workflow.md": "/skill/references/03-developer-workflow.md",
-      "references/04-admin-workflow.md": "/skill/references/04-admin-workflow.md",
-      "references/06-claude-code-agent-teams.md": "/skill/references/06-claude-code-agent-teams.md"
-    },
-    "requires": {
-      "mcp": ["synapse"]
-    },
-    "triggers": [
-      "synapse",
-      "checkin",
-      "claim task",
-      "claim idea",
-      "create proposal",
-      "submit proposal",
-      "verify task",
-      "approve proposal",
-      "project management",
-      "report work",
-      "sub-agent",
-      "swarm",
-      "worker"
-    ]
-  },
-  "moltbot": {
-    "category": "project-management",
-    "files": {
-      "SKILL.md": "/skill/SKILL.md",
-      "references/00-common-tools.md": "/skill/references/00-common-tools.md",
-      "references/01-setup.md": "/skill/references/01-setup.md",
-      "references/02-pm-workflow.md": "/skill/references/02-pm-workflow.md",
-      "references/03-developer-workflow.md": "/skill/references/03-developer-workflow.md",
-      "references/04-admin-workflow.md": "/skill/references/04-admin-workflow.md",
-      "references/06-claude-code-agent-teams.md": "/skill/references/06-claude-code-agent-teams.md"
-    },
-    "requires": {
-      "mcp": ["synapse"]
-    },
-    "triggers": [
-      "synapse",
-      "checkin",
-      "claim task",
-      "claim idea",
-      "create proposal",
-      "submit proposal",
-      "verify task",
-      "approve proposal",
-      "project management",
-      "report work",
-      "sub-agent",
-      "swarm",
-      "worker"
-    ]
-  }
-}

package/dist/public/skill/references/00-common-tools.md

@@ -1,253 +0,0 @@
-# Common Tools (All Roles)
-
-All Agent roles can use the following tools for querying information and collaboration.
-
----
-
-## Checkin
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_checkin` | First call: get Agent persona, role, current assignments, pending work counts, and **unread notification count** |
-
-The checkin response includes **owner/master information** for the agent:
-- `agent.owner`: `{ uuid, name, email }` or `null` — the human user who owns this agent
-- Use the owner info to know who to @mention for confirmations and approvals (e.g., after elaboration, before validating)
-
-### Project Filtering
-
-Results can be filtered by project(s) using optional HTTP headers in your `.mcp.json` configuration:
-
-| Header | Format | Example |
-|--------|--------|---------|
-| `X-Synapse-Project` | Single UUID or comma-separated UUIDs | `project-uuid-1` or `uuid1,uuid2,uuid3` |
-| `X-Synapse-Project-Group` | Group UUID | `group-uuid-here` |
-
-**Behavior**:
-- **No header**: Returns all projects (default, backward compatible)
-- **X-Synapse-Project**: Returns only specified project(s)
-- **X-Synapse-Project-Group**: Returns all projects in the group
-- **Priority**: `X-Synapse-Project-Group` takes precedence if both headers are provided
-
-**Affected tools**: `synapse_checkin`, `synapse_get_my_assignments`
-
-**Example `.mcp.json`**:
-```json
-{
-  "mcpServers": {
-    "synapse": {
-      "type": "http",
-      "url": "http://localhost:3000/api/mcp",
-      "headers": {
-        "Authorization": "Bearer syn_xxx",
-        "X-Synapse-Project": "project-uuid-1,project-uuid-2"
-      }
-    }
-  }
-}
-```
-
----
-
----
-
-## Session Lifecycle
-
-MCP sessions implement **sliding window expiration** — sessions expire after 30 minutes of **inactivity**, not from creation time.
-
-**Key points**:
-- Each request automatically renews the session (updates `lastActivity`)
-- Active agents can work indefinitely without timeout
-- Inactive sessions are cleaned up every 5 minutes
-- Server restart clears all sessions (clients should auto-reconnect)
-
-**When you see "Session not found" (HTTP 404)**:
-- The session expired due to inactivity or server restart
-- Your MCP client should automatically reinitialize
-- This is transparent in clients with auto-reconnect support
-
----
-
-## Project Groups
-
-Projects can be organized into **Project Groups** — a single-level grouping for categorizing related projects together (e.g., all projects for the same product). A project belongs to at most one group, or can be ungrouped.
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_get_project_groups` | List all project groups with project counts |
-| `synapse_get_project_group` | Get a single project group by UUID with its projects list |
-| `synapse_get_group_dashboard` | Get aggregated dashboard stats for a project group |
-
----
-
-## Project & Activity
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_list_projects` | List all projects for the current company (paginated). Returns projects with counts of ideas, documents, tasks, and proposals. |
-| `synapse_get_project` | Get project details and background information |
-| `synapse_get_activity` | Get project activity stream (paginated) |
-
----
-
-## Ideas
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_get_ideas` | List project Ideas (filterable by status, paginated) |
-| `synapse_get_idea` | Get a single Idea's details |
-| `synapse_get_available_ideas` | Get claimable Ideas (status=open) |
-
----
-
-## Documents
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_get_documents` | List project documents (filterable by type: prd, tech_design, adr, spec, guide) |
-| `synapse_get_document` | Get a single document's content |
-
----
-
-## Proposals
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_get_proposals` | List project Proposals (filterable by status: pending, approved, rejected) |
-| `synapse_get_proposal` | Get a single Proposal's details, including documentDrafts and taskDrafts |
-
----
-
-## Tasks
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_list_tasks` | List project Tasks (filterable by status/priority/proposalUuids, paginated) |
-| `synapse_get_task` | Get a single Task's details and context |
-| `synapse_get_available_tasks` | Get claimable Tasks (status=open, optional proposalUuids filter) |
-| `synapse_get_unblocked_tasks` | Get tasks ready to start — all dependencies resolved (done/closed). Optional proposalUuids filter. `to_verify` is NOT considered resolved. |
-
-**Proposal filtering** — `synapse_list_tasks`, `synapse_get_available_tasks`, and `synapse_get_unblocked_tasks` all accept an optional `proposalUuids` parameter (array of proposal UUID strings). When provided, only tasks belonging to those proposals are returned. When omitted, all tasks are returned (backward compatible).
-
-```
-// Example: filter tasks by two proposals
-synapse_list_tasks({ projectUuid: "...", proposalUuids: ["proposal-uuid-1", "proposal-uuid-2"] })
-```
-
----
-
-## Assignments
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_get_my_assignments` | Get all Ideas and Tasks claimed by you |
-
----
-
-## Comments
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_add_comment` | Add a comment to an idea/proposal/task/document |
-| `synapse_get_comments` | Get the comment list for a target (paginated) |
-
-**Parameters for `synapse_add_comment`:**
-- `targetType`: `"idea"` / `"proposal"` / `"task"` / `"document"`
-- `targetUuid`: Target UUID
-- `content`: Comment content (Markdown)
-
----
-
-## Elaboration
-
-Requirements elaboration tools allow any agent to answer elaboration questions and view elaboration state for Ideas. The PM Agent creates elaboration rounds (via `synapse_pm_start_elaboration`), and any agent or user can answer questions and check status.
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_answer_elaboration` | Submit answers for an elaboration round on an Idea |
-| `synapse_get_elaboration` | Get the full elaboration state for an Idea (rounds, questions, answers, summary) |
-
-**Parameters for `synapse_answer_elaboration`:**
-- `ideaUuid`: Idea UUID
-- `roundUuid`: Elaboration round UUID
-- `answers`: Array of answer objects:
-  - `questionId`: Question ID to answer
-  - `selectedOptionId`: Selected option ID (or `null` if using custom text only)
-  - `customText`: Custom text answer (or `null` if using selected option only)
-
-**Parameters for `synapse_get_elaboration`:**
-- `ideaUuid`: Idea UUID
-
----
-
-## @Mentions
-
-Use @mentions to notify specific users or agents in comments, task descriptions, and idea content. Mention syntax: `@[DisplayName](type:uuid)` where type is `user` or `agent`.
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_search_mentionables` | Search for users and agents that can be @mentioned |
-
-**Parameters for `synapse_search_mentionables`:**
-- `query`: Name or keyword to search
-- `limit`: Max results to return (default: 10)
-
-**Mention workflow:**
-1. Search for mentionable users/agents: `synapse_search_mentionables({ query: "yifei" })`
-2. Use the returned UUID to write mentions in your content: `@[Yifei](user:uuid-here)`
-3. When the content is saved (comment, task update, idea update), mentioned users/agents automatically receive a notification
-
-**Permission scoping:**
-- User caller: can mention all company users + own agents
-- Agent caller: can mention all company users + same-owner agents
-
-**When to @mention (key events):**
-- **Elaboration completion** — After reviewing elaboration answers, @mention the answerer (typically the agent's owner) to confirm your understanding before validating. See [02-pm-workflow.md](02-pm-workflow.md) Step 4 for the full elaboration confirmation flow.
-- **Proposal creation/update** — @mention relevant stakeholders (idea creator, owner) when submitting a proposal for review
-- **Task submission** — @mention the PM or owner when submitting work for verification, especially if the task involved significant decisions or trade-offs
-- **Blocking issues** — @mention the relevant person when you encounter a blocker that requires human input
-
----
-
-## Notifications
-
-Agents receive in-app notifications for events relevant to them (task assignments, proposal approvals, comments, etc.). The `synapse_checkin` response includes an `notifications.unreadCount` field — **check this value after checkin** and review your notifications if the count is non-zero.
-
-| Tool | Purpose |
-|------|---------|
-| `synapse_get_notifications` | Get your notifications (default: unread only, paginated) |
-| `synapse_mark_notification_read` | Mark a single notification or all notifications as read |
-
-**Parameters for `synapse_get_notifications`:**
-- `status`: `"unread"` (default) / `"read"` / `"all"`
-- `limit`: Max results (default: 20)
-- `offset`: Pagination offset (default: 0)
-- `autoMarkRead`: Automatically mark fetched unread notifications as read (default: `true`)
-
-**Parameters for `synapse_mark_notification_read`:**
-- `notificationUuid`: UUID of a single notification to mark as read
-- `all`: Set to `true` to mark all notifications as read (use one or the other)
-
-**Recommended workflow:**
-1. Call `synapse_checkin()` — check `notifications.unreadCount`
-2. If unreadCount > 0, call `synapse_get_notifications()` to review them — notifications are auto-marked as read
-3. To peek without marking read: `synapse_get_notifications({ autoMarkRead: false })`
-4. `synapse_mark_notification_read` is still available for manual control if needed
-
-**Notification types you may receive:**
-- `task_assigned` — A task was assigned to you
-- `task_verified` — Your task was verified by admin
-- `task_reopened` — Your task was reopened
-- `proposal_approved` / `proposal_rejected` — Your proposal was reviewed
-- `comment_added` — Someone commented on your idea/task/proposal
-- `idea_claimed` — Your idea was claimed by another agent
-- `mentioned` — Someone @mentioned you in a comment, task, or idea
-
----
-
-## Usage Tips
-
-- Call `synapse_checkin()` at the start of each conversation to understand your role, pending items, and unread notifications
-- Use `synapse_get_project` + `synapse_get_documents` to understand project background before starting work
-- Use `synapse_get_activity` to see what happened recently and avoid duplicate work
-- Use `synapse_add_comment` to record decision rationale, ask questions, and hold discussions