@@ -22,6 +22,9 @@ This file tracks the current deterministic runtime baseline for Loader. It stays |
| 22 | 22 | - durable project memory in `.loader/project-memory.json` and working notes in `.loader/notepad.md` |
| 23 | 23 | - native memory tools for `project_memory_*` and `notepad_*` |
| 24 | 24 | - heuristic workflow routing across `clarify` → `plan` → `execute` → `verify` |
| 25 | +- clarify mode as an explicit single-question brief flow that returns to execute mode |
| 26 | +- plan mode as explicit single-pass implementation and verification artifact generation |
| 27 | +- persisted workflow-artifact status and source metadata in session state when execute consumes or reuses workflow artifacts |
| 25 | 28 | - mode-specific system prompts for clarify, plan, execute, and verify |
| 26 | 29 | - explicit verify/fix loops for mutating tasks, with a bounded retry budget |
| 27 | 30 | - verify/fix retries return to execute mode without re-triggering clarify or plan |
@@ -47,10 +50,10 @@ This file tracks the current deterministic runtime baseline for Loader. It stays |
| 47 | 50 | ## Known weak spots |
| 48 | 51 | |
| 49 | 52 | - the core turn loop moved into [`src/loader/runtime/conversation.py`](../src/loader/runtime/conversation.py), but it still owns workflow routing, remaining loop safeguards, and other coordination logic that remains more heuristic-heavy than the reference runtime in `refs/claw-code` |
| 50 | | -- planning, decomposition, and several helper behaviors still live in [`src/loader/agent/loop.py`](../src/loader/agent/loop.py), so ownership is cleaner than Sprint 00 but not fully simplified yet |
| 53 | +- workflow routing is cleaner than Sprint 00, but the router and artifact bridge still live in [`src/loader/runtime/conversation.py`](../src/loader/runtime/conversation.py) and remain more heuristic than the reference runtimes |
| 51 | 54 | - the mode router is still heuristic-only; Loader does not yet implement OMX's deeper ambiguity scoring, pressure-pass discipline, or branch-specific routing policy |
| 52 | | -- clarify mode currently stops after one structured question and one brief artifact; it does not yet run a deeper Socratic loop |
| 53 | | -- plan mode is still a single-pass artifact generator, not a Planner/Architect/Critic consensus loop |
| 55 | +- clarify mode is now explicitly a single-question brief flow, not a deeper Socratic protocol |
| 56 | +- plan mode is now explicitly a single-pass artifact generator, not a Planner/Architect/Critic consensus loop |
| 54 | 57 | - DoD acceptance criteria and pending items are stronger than Sprint 02, but todo progress is still lightly structured compared with claw-code's richer workflow state |
| 55 | 58 | - evidence summaries are deterministic runtime summaries of captured output, not model-written verification narratives |
| 56 | 59 | - session compaction summaries are heuristic runtime summaries, not model-assisted continuity artifacts |
@@ -109,13 +112,13 @@ The auditable manifest lives at [`tests/fixtures/runtime_parity_manifest.json`]( |
| 109 | 112 | |
| 110 | 113 | As of 2026-04-07: |
| 111 | 114 | |
| 112 | | -- `uv run pytest -q`: 210 passed |
| 115 | +- `uv run pytest -q`: 211 passed |
| 113 | 116 | - `tests/test_runtime_harness.py` is fully green, including permission-mode parity, DoD verify/fix coverage, workflow routing parity, and the original contract regression |
| 114 | 117 | - `tests/test_dod.py` covers persistence, sizing boundaries, and verification command derivation |
| 115 | 118 | - `tests/test_workflow.py` covers router heuristics, clarify/plan artifact round trips, DoD workflow links, and todo-to-DoD syncing |
| 116 | 119 | - `tests/test_workflow_runtime.py` covers clarify routing, plan routing, and verify-fix workflow handoff |
| 117 | 120 | - `tests/test_workflow_tools.py` and `tests/test_workflow_runtime_tools.py` cover `TodoWrite`, `AskUserQuestion`, and runtime callback plumbing |
| 118 | | -- `tests/test_session_state.py` covers session persistence, resume, rotation, compaction persistence, cumulative usage rollups, and persisted permission-policy metadata |
| 121 | +- `tests/test_session_state.py` covers session persistence, resume, rotation, compaction persistence, cumulative usage rollups, persisted permission-policy metadata, and persisted workflow-artifact state |
| 119 | 122 | - `tests/test_compaction.py` covers claw-style line compression and compacted continuation-message behavior |
| 120 | 123 | - `tests/test_memory_tools.py` covers project-memory writes, notepad writes, lifecycle-hook mirroring, and DoD-summary capture into project memory |
| 121 | 124 | - `tests/test_cli_resume.py` covers `--resume` argument rewriting for latest and named-session restore |
@@ -135,7 +138,7 @@ As of 2026-04-07: |
| 135 | 138 | - Sprint 01 turned the original `tool_call_id` regression green by fixing the message contract, not by weakening the test. |
| 136 | 139 | - Sprint 02 replaced "looks done" completion for mutating tasks with a real verify/fix gate, but it has not yet reached the richer workflow contracts described in the report and Sprint 04+. |
| 137 | 140 | - Sprint 03 established permission modes, hooks, and tool hardening, but it intentionally stops short of claw-code's fuller rule engine and prompt/allow permission variants. |
| 138 | | -- Sprint 04 adds routing, artifacts, and structured user questions, but it is still a first-pass workflow layer rather than full OMX consensus planning or deep interview rigor. |
| 141 | +- Sprint 04's workflow layer is now explicitly scoped as lightweight: single-question clarify, single-pass planning, explicit artifact bridging, and no legacy decomposition path. |
| 139 | 142 | - Sprint 05 adds durable sessions, resume, compaction, and native memory/notepad tools, but it stops short of Sprint 06's inspectable session/status product surfaces and still uses heuristic continuity summaries rather than richer semantic memory extraction. |
| 140 | 143 | - Sprint 06 adds inspectable product surfaces, a constrained explore lane, and a broader tool registry, but it still stops short of interactive explore workflows, richer git ergonomics, AST/LSP-aware editing, or any multi-agent/team runtime. |
| 141 | 144 | - Sprint 07 is complete: Loader now has prompt/allow modes, rule-based permission policy, policy-backed prompting, persisted policy inspection state, and smaller assistant-turn/tool-batch/finalization runtime seams, but it still stops short of a richer rule UX, deeper policy sandboxing, and the more opinionated workflow/runtime contracts in the refs. |
