diff --git a/README.md b/README.md
index ff4413b20f..6d41f0e657 100644
--- a/README.md
+++ b/README.md
@@ -105,6 +105,8 @@ Human gate        Human gate                              Human merge
 3. **Implement**: The plan's tasks are split into a **DAG of independent slices** — each slice runs as its own agent team on its own integration branch with its own BRC consensus and stacked PR. Slices whose dependencies are satisfied run concurrently (up to `EGG_ORCH_MAX_PARALLEL_SLICES`, default 2 per pipeline; a separate `EGG_ORCH_GLOBAL_MAX_PARALLEL_SLICES` cap, default 4, bounds the total across all running pipelines); slices with unmet dependencies wait in subsequent waves. Within each slice, the coder writes code, the tester writes regression tests and adversarially probes the coder's implementation for bugs (NACKing with failing tests as bug reports), and the documenter updates docs. Code and contract reviewers provide line-level feedback; security and concurrency lens reviewers add targeted cross-file analysis and block consensus on a NACK. Cycles continue until all checks pass and BRC consensus is reached for that slice. See [Slice-DAG Implement Phase](docs/architecture/slice-dag.md) for the full model.
 4. **PR**: Orchestrator auto-creates the PR from plan metadata. Only a human can merge via GitHub UI.
 
+For **Jira epic-mode pipelines** (when `jira_ticket` resolves to a Jira Epic), an **Apply** phase is inserted between Plan and Implement: the `applier` role drives Jira mutations (epic Description writes, child ticket creates/edits, link creates, Won't-Do handoffs) on operator approval, before code implementation begins. Pass `mode='fresh'` (no existing children) or `mode='reassess'` (existing children to classify) to `submit_task`; `mode='auto'` (default) detects which to use at submit time.
+
 Within each phase, specialized agents run concurrently via BRC (enabled by default for refine, plan, and implement). Here's what a completed pipeline looks like:
 
 ```
diff --git a/docs/reference/agent-roles.md b/docs/reference/agent-roles.md
index aac75a7a69..0e3eae8919 100644
--- a/docs/reference/agent-roles.md
+++ b/docs/reference/agent-roles.md
@@ -8,7 +8,7 @@ Every agent role belongs to one of five categories. Categories enable dynamic te
 
 | Category | Purpose | Roles |
 |----------|---------|-------|
-| **EXECUTION** | Produce artifacts (code, tests, docs) | `coder`, `tester`, `documenter` |
+| **EXECUTION** | Produce artifacts (code, tests, docs, Jira mutations) | `coder`, `tester`, `documenter`, `applier` |
 | **ANALYSIS** | Analyze tasks and plan work | `refiner`, `architect`, `task_planner`, `risk_analyst` |
 | **REVIEW** | Validate quality and correctness | `reviewer_code`, `reviewer_code_holistic`, `reviewer_contract`, `reviewer_refine`, `reviewer_plan`, `reviewer_agent_design`, `reviewer_security`, `reviewer_concurrency` |
 | **UTILITY** | Cross-cutting support tasks | `autofixer`, `conflict_resolver` |
@@ -27,6 +27,7 @@ Use `get_roles_by_category(AgentCategory.REVIEW)` to dynamically query roles by
 | `task_planner` | Analysis | Plan | Yes (with `risk_analyst`) | architect |
 | `risk_analyst` | Analysis | Plan | Yes (with `task_planner`) | architect |
 | `reviewer_plan` | Review | Plan | No | task_planner, risk_analyst |
+| `applier` | Execution | Apply (epic-mode only) | No | — |
 | `coder` | Execution | Implement | No | — |
 | `tester` | Execution | Implement | Yes (with `documenter`) | coder |
 | `documenter` | Execution | Implement | Yes (with `tester`) | coder |
@@ -130,6 +131,24 @@ All agents within a phase run concurrently via BRC consensus. Concurrency is ena
 **Outputs**:
 - `.egg-state/reviews/{identifier}-plan-plan-review.json` — Verdict file
 
+## Apply Phase
+
+The apply phase is **conditional** — it is only inserted between Plan and Implement when `Pipeline.is_epic` is true (i.e., the submitted `jira_ticket` resolves to a Jira Epic). Non-epic pipelines advance directly from Plan to Implement.
+
+### `applier`
+
+**Purpose**: Drive Jira mutations on operator approval of the refine/plan HITL gates for epic-mode pipelines. For each task in the contract, the applier dispatches all `jira_action` values *except* `wontdo` (i.e., `create`, `edit`, `split-of`, `consolidate-into`); `wontdo` is handled via the handoff JSON described below — the applier emits the handoff entry, and the orchestrator drains it via the orchestrator-only `/api/v1/jira/ticket/transition` route.
+
+**File access**:
+- Allowed writes: `.egg-state/agent-outputs/`
+- Blocked: all source code, docs, tests, contracts, drafts, `.github/`
+
+**Outputs**:
+- `.egg-state/agent-outputs/{pipeline_id}-applier-output.json` — Per-action apply results (`jira_action_status`: `applied` / `failed`)
+- `.egg-state/agent-outputs/{pipeline_id}-wontdo.json` — Handoff JSON listing tickets to transition to "Won't Do"; drained by the orchestrator after APPLY consensus
+
+**Prompt context**: `EGG_IS_EPIC`, `EGG_EPIC_MODE` (`epic-fresh` or `epic-reassess`), contract path, plan draft.
+
 ## Implement Phase
 
 The three producer roles in this phase — `coder`, `tester`, and `documenter` —