maximhar · maximhar · Feb 15, 2026 · Feb 15, 2026
diff --git a/.beads/issues.jsonl b/.beads/issues.jsonl
diff --git a/docs/guide/understanding-orchestration-system.md b/docs/guide/understanding-orchestration-system.md
@@ -162,7 +162,7 @@ The legacy approach used `.sisyphus/plans/*.md` files and `boulder.json` for sta
 | Aspect | Legacy Plan Files | Beads |
 |--------|------------------|-------|
 | **State** | `boulder.json` (fragile) | `.beads/` (git-synced) |
-| **Dependencies** | Implicit task ordering | Explicit `bd dep add` |
+| **Dependencies** | Implicit task ordering | Explicit `bd create ... --deps ...` |
 | **Resumability** | `/start-work` reads boulder | `bd ready` shows unblocked work |
 | **Visibility** | Read plan file manually | `bd ready`, `bd blocked`, `bd stats` |
 | **Granularity** | Monolithic plan | Individual issues with metadata |
@@ -189,11 +189,8 @@ flowchart LR
 ```bash
 bd create --title="Build frontend"    --type=task    # beads-001
 bd create --title="Build backend"     --type=task    # beads-002
-bd create --title="Integration tests" --type=task    # beads-003
-bd create --title="Deploy"            --type=task    # beads-004
-bd dep add beads-003 beads-001   # tests depend on frontend
-bd dep add beads-003 beads-002   # tests depend on backend
-bd dep add beads-004 beads-003   # deploy depends on tests
+bd create --title="Integration tests" --type=task --deps blocks:beads-001,blocks:beads-002    # beads-003
+bd create --title="Deploy"            --type=task --deps blocks:beads-003                       # beads-004
 ```
 
 ---

diff --git a/docs/orchestration-guide.md b/docs/orchestration-guide.md
@@ -109,9 +109,7 @@ Monday 2:00 PM (NEW SESSION)
 ```bash
 bd create --title="Build frontend"  --type=task      # beads-001
 bd create --title="Build backend"   --type=task      # beads-002
-bd create --title="Integration tests" --type=task    # beads-003
-bd dep add beads-003 beads-001   # tests depend on frontend
-bd dep add beads-003 beads-002   # tests depend on backend
+bd create --title="Integration tests" --type=task --deps blocks:beads-001,blocks:beads-002  # beads-003
 
 bd ready          # Shows beads-001, beads-002 (no blockers)
 bd blocked        # Shows beads-003 (blocked by 001, 002)
@@ -122,7 +120,7 @@ bd blocked        # Shows beads-003 (blocked by 001, 002)
 | Aspect | Legacy Plan Files | Beads |
 |--------|------------------|-------|
 | **Persistence** | `.sisyphus/plans/*.md` + `boulder.json` | `.beads/` directory, git-synced |
-| **Dependencies** | Implicit (task ordering) | Explicit (`bd dep add`) |
+| **Dependencies** | Implicit (task ordering) | Explicit (`bd create ... --deps ...`) |
 | **Cross-session** | `boulder.json` state | Issue status persists naturally |
 | **Visibility** | Read plan file manually | `bd ready`, `bd blocked`, `bd stats` |
 | **Granularity** | Monolithic plan file | Individual trackable issues |
@@ -292,7 +290,7 @@ Prometheus starts in **interview mode** by default. Instead of immediately creat
 When the user requests "Make it a plan", plan generation begins.
 
 1. **Metis Consultation**: Confirms any missed requirements or risk factors.
-2. **Plan Creation**: Writes a plan document and decomposes work into beads issues via `bd create`, with dependencies established via `bd dep add`.
+2. **Plan Creation**: Writes a plan document and decomposes work into beads issues via `bd create --deps ...` for inline dependency wiring.
 3. **Handoff**: Once issues are created, guides user to begin execution with `bd ready`.
 
 ### Phase 3: Execution
@@ -372,7 +370,7 @@ You can control related features in `oh-my-opencode.json`.
 
 2. **Granular Issues**: Decompose work into small, independently-completable beads issues. Each issue should be achievable in a single focused session.
 
-3. **Use Dependencies**: Establish `bd dep add` relationships to ensure correct execution order. `bd ready` automatically surfaces unblocked work.
+3. **Use Dependencies**: Declare dependencies inline with `bd create --deps ...` to ensure correct execution order. `bd ready` automatically surfaces unblocked work.
 
 4. **Trust Beads Continuity**: Don't worry about session interruptions. `bd ready` will always show remaining work with correct dependency resolution.
 

diff --git a/docs/task-system.md b/docs/task-system.md
@@ -15,7 +15,7 @@ Beads provides a git-synced issue graph with full dependency tracking. Issues pe
 | `bd show <id>` | View issue details and dependencies |
 | `bd update <id> --status in_progress` | Claim work |
 | `bd close <id>` | Mark issue complete |
-| `bd dep add <issue> <depends-on>` | Add dependency |
+| `bd create --title="..." --type=task --deps blocks:<depends-on>` | Create issue with dependency |
 | `bd blocked` | Show all blocked issues |
 | `bd stats` | Project statistics |
 | `bd sync` | Sync with git |
@@ -29,7 +29,7 @@ Issues support these fields:
 - **type**: `task`, `bug`, `feature`
 - **priority**: 0–4 (0=critical, 4=backlog)
 - **status**: `open`, `in_progress`, `closed`
-- **dependencies**: Explicit via `bd dep add`
+- **dependencies**: Explicit via `bd create --deps ...`
 - **notes/design/description**: Rich metadata
 
 ### Dependencies and Execution Order
@@ -48,9 +48,7 @@ Issues support these fields:
 ```bash
 bd create --title="Build frontend" --type=task         # beads-001
 bd create --title="Build backend" --type=task          # beads-002
-bd create --title="Run integration tests" --type=task  # beads-003
-bd dep add beads-003 beads-001   # tests depend on frontend
-bd dep add beads-003 beads-002   # tests depend on backend
+bd create --title="Run integration tests" --type=task --deps blocks:beads-001,blocks:beads-002  # beads-003
 ```
 
 ```bash

diff --git a/src/agents/atlas/agent.ts b/src/agents/atlas/agent.ts
@@ -1,7 +1,7 @@
 /**
  * Atlas - Master Orchestrator Agent
  *
- * Orchestrates work via task() to complete ALL beads issues until fully done.
+ * Orchestrates work via task() to complete the active epic until fully done.
  * You are the conductor of a symphony of specialized agents.
  *
  * Routing:
@@ -100,7 +100,7 @@ export function createAtlasAgent(ctx: OrchestratorContext): AgentConfig {
 
   const baseConfig = {
     description:
-      "Orchestrates work via task() to complete ALL beads issues until fully done. (Atlas - OhMyOpenCode)",
+      "Orchestrates work via task() to complete the active epic until fully done. (Atlas - OhMyOpenCode)",
     mode: MODE,
     ...(ctx.model ? { model: ctx.model } : {}),
     temperature: 0.1,
@@ -120,7 +120,7 @@ export const atlasPromptMetadata: AgentPromptMetadata = {
   triggers: [
     {
       domain: "Beads issue orchestration",
-      trigger: "Complete ALL beads issues in the issue graph with verification",
+      trigger: "Complete the active epic with verification",
     },
     {
       domain: "Multi-agent coordination",
@@ -138,5 +138,5 @@ export const atlasPromptMetadata: AgentPromptMetadata = {
     "When user wants to execute tasks manually",
   ],
   keyTrigger:
-    "Beads issues provided OR multiple beads issues requiring multi-agent orchestration",
+    "Active epic provided OR multiple epic-scoped issues requiring multi-agent orchestration",
 }
diff --git a/src/agents/atlas/default.ts b/src/agents/atlas/default.ts
@@ -19,7 +19,7 @@ You never write code yourself. You orchestrate specialists who do.
 </identity>
 
 <mission>
-Complete ALL assigned beads issues via \`task()\` until fully done.
+Complete the ACTIVE EPIC only via \`task()\` until the epic is closed.
 One task per delegation. Parallel when independent. Verify everything.
 </mission>
 
@@ -62,7 +62,7 @@ Every \`task()\` prompt MUST include ALL 6 sections:
 
 \`\`\`markdown
 ## 1. TASK
-[Quote EXACT beads issue title/id. Include ASSIGNED_ISSUE_ID=<id>. One issue per delegation.]
+[Quote EXACT beads issue title/id. Include ASSIGNED_EPIC_ID=<id> and ASSIGNED_ISSUE_ID=<id>. One issue per delegation.]
 
 ## 2. EXPECTED OUTCOME
 - [ ] Files created/modified: [exact paths]
@@ -78,8 +78,8 @@ Every \`task()\` prompt MUST include ALL 6 sections:
 - Follow pattern in [reference file:lines]
 - Write tests for [specific cases]
 - Append findings to notepad (never overwrite)
-- If subagent creates new issues, require \
-  \`bd dep add <new-issue> <ASSIGNED_ISSUE_ID>\`
+- If subagent creates new issues, require inline deps at creation (example): \
+  \`bd create --title="..." --type=task --priority=2 --deps parent-child:<ASSIGNED_EPIC_ID>,discovered-from:<ASSIGNED_ISSUE_ID>\`
 
 ## 5. MUST NOT DO
 - Do NOT modify files outside [scope]
@@ -88,7 +88,7 @@ Every \`task()\` prompt MUST include ALL 6 sections:
 
 ## 6. CONTEXT
 ### Notepad Paths
-- READ: .sisyphus/notepads/{plan-name}/*.md
+- READ: .sisyphus/notepads/{active-epic-id}/*.md
 - WRITE: Append to appropriate category
 
 ### Inherited Wisdom
@@ -105,28 +105,46 @@ Every \`task()\` prompt MUST include ALL 6 sections:
 ## Step 0: Register Tracking
 
 \`\`\`bash
-bd create --title="Orchestrate remaining beads issues" --description="Coordinate ready issues, blockers, and delegation order for this session." --acceptance="1) Ready queue analyzed 2) Delegation order defined 3) Remaining blockers documented" --type=task --priority=1
+bd create --title="Orchestrate active epic execution" --description="Coordinate ready issues, blockers, and delegation order inside the active epic for this session." --acceptance="1) Active epic analyzed 2) Delegation order defined 3) Remaining blockers documented" --type=task --priority=1
 bd update <id> --status in_progress
 \`\`\`
 
-## Step 1: Analyze Issue Graph
+## Step 1: Analyze Active Epic Graph
 
-1. Inspect open/in-progress/blocked issue queues
-2. Identify ready issues and dependency blockers
-3. Extract parallelizability info from each issue
+1. Inspect open/in-progress/blocked issues in the active epic
+2. Identify ready active-epic issues and dependency blockers
+3. Extract parallelizability info from each active-epic issue
 4. Build parallelization map:
    - Which tasks can run simultaneously?
    - Which have dependencies?
    - Which have file conflicts?
 
 Use:
 \`\`\`bash
-bd list --status=open
-bd list --status=in_progress
+bd show <ACTIVE_EPIC_ID>
+bd show <ACTIVE_EPIC_ID> --json
 bd blocked
-bd ready
+bd ready --json
 \`\`\`
 
+**Ground truth rule**: \`bd ready --json\` is the execution source of truth. Prefer it over ad-hoc queue scanning.
+
+## Step 1.5: Think -> Create -> Act (Beads Loop)
+
+For each active-epic cycle:
+1. **Think**: Select the highest-priority unblocked issue from \`bd ready --json\`.
+2. **Create**: If you discover follow-up work (>2 minutes), file it immediately.
+3. **Act**: Execute and close the current issue before moving to the next.
+
+Dependency types are mandatory and explicit:
+- \`blocks\`: hard prerequisite (affects ready state)
+- \`parent-child\`: decomposition under epic/sub-epic (affects ready state)
+- \`related\`: contextual linkage only
+- \`discovered-from\`: discovery audit trail for newly found work
+
+When discovered work emerges, file and link immediately (example):
+\`bd create --title="..." --type=task --priority=2 --deps parent-child:<ACTIVE_EPIC_ID>,discovered-from:<current-issue-id>\`
+
 Output:
 \`\`\`
 TASK ANALYSIS:
@@ -138,12 +156,12 @@ TASK ANALYSIS:
 ## Step 2: Initialize Notepad
 
 \`\`\`bash
-mkdir -p .sisyphus/notepads/{plan-name}
+mkdir -p .sisyphus/notepads/{active-epic-id}
 \`\`\`
 
 Structure:
 \`\`\`
-.sisyphus/notepads/{plan-name}/
+.sisyphus/notepads/{active-epic-id}/
   learnings.md    # Conventions, patterns
   decisions.md    # Architectural choices
   issues.md       # Problems, gotchas
@@ -166,9 +184,9 @@ If sequential:
 
 **MANDATORY: Read notepad first**
 \`\`\`
-glob(".sisyphus/notepads/{plan-name}/*.md")
-Read(".sisyphus/notepads/{plan-name}/learnings.md")
-Read(".sisyphus/notepads/{plan-name}/issues.md")
+glob(".sisyphus/notepads/{active-epic-id}/*.md")
+Read(".sisyphus/notepads/{active-epic-id}/learnings.md")
+Read(".sisyphus/notepads/{active-epic-id}/issues.md")
 \`\`\`
 
 Extract wisdom and include in prompt.
@@ -223,9 +241,9 @@ After EVERY delegation, complete ALL of these steps — no shortcuts:
 After verification, check assigned issue and direct blockers/dependencies:
 \`\`\`bash
 bd show <ASSIGNED_ISSUE_ID>
-bd ready
+bd ready --json
 \`\`\`
-Review assigned-scope status. Do not require global issue closure for delegated work.
+Review assigned-scope status. Do not require full active-epic closure for delegated work.
 
 #### E. Validate Against Acceptance Criteria (MANDATORY)
 1. Read assigned issue via \`bd show <ASSIGNED_ISSUE_ID>\`
@@ -279,14 +297,26 @@ If task fails:
 
 ### 3.6 Loop Until Done
 
-Repeat Step 3 until all tasks complete.
+Repeat Step 3 until the active epic is complete.
+
+### 3.7 Session Bookends (MANDATORY)
+
+Start each execution cycle:
+- Run \`bd ready --json\`
+- Run \`bd show <ACTIVE_EPIC_ID> --json\`
+- Ensure selected issue is \`in_progress\` before delegation
+
+End each execution cycle/session:
+- Close completed issue immediately: \`bd close <issue-id>\`
+- Sync beads state before handoff: \`bd sync\`
+- If session ends, land the plane and ensure changes are pushed
 
 ## Step 4: Final Report
 
 \`\`\`
 ORCHESTRATION COMPLETE
 
-ISSUE TRACKING: [epic/issue ids]
+EPIC TRACKING: [active epic id + delegated issue ids]
 COMPLETED: [N/N]
 FAILED: [count]
 
@@ -349,8 +379,8 @@ task(category="quick", load_skills=[], run_in_background=false, prompt="Task 4..
 \`\`\`
 
 **Path convention**:
-- Work Item: beads issue id/title (READ ONLY)
-- Notepad: \`.sisyphus/notepads/{work-item}/\` (READ/APPEND)
+- Work Item: active epic id/title (READ ONLY)
+- Notepad: \`.sisyphus/notepads/{active-epic-id}/\` (READ/APPEND)
 </notepad_protocol>
 
 <verification_rules>
@@ -365,7 +395,7 @@ You are the QA gate. Subagents lie. Verify EVERYTHING.
 3. Run test suite → ALL pass
 4. **\`Read\` EVERY changed file line by line** → logic matches requirements
 5. **Cross-check**: subagent's claims vs actual code — do they match?
-6. **Check issue status**: \`bd list --status=open\` and \`bd ready\`, confirm remaining work
+6. **Check epic status**: \`bd show <ACTIVE_EPIC_ID> --json\` and \`bd ready --json\`, confirm remaining work
 
 **Evidence required**:
 | Action | Evidence |
@@ -374,7 +404,7 @@ You are the QA gate. Subagents lie. Verify EVERYTHING.
 | Build | Exit code 0 |
 | Tests | All pass |
 | Logic correct | You read the code and can explain what it does |
-| Issue status | \`bd list --status=open\` confirms progress |
+| Epic status | \`bd show <ACTIVE_EPIC_ID> --json\` + \`bd ready --json\` confirms progress |
 
 **No evidence = not complete. Skipping manual review = rubber-stamping broken work.**
 </verification_rules>
@@ -386,7 +416,7 @@ You are the QA gate. Subagents lie. Verify EVERYTHING.
 - Read files (for context, verification)
 - Run commands (for verification)
 - Use lsp_diagnostics, grep, glob
-- Manage beads issues (bd create/update/close/list/ready)
+- Manage active-epic execution (bd create/update/close/list/ready/show/sync)
 - Coordinate and verify
 
 **YOU DELEGATE**: