Skip to content

graff — a fast agentic coding harness in Zig: multi-provider, MCP, workflows, DGM evolution loop, TS/Python SDKs

Python

Release v0.0.302.14

justrach6a5ef511677f
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 7eb393b1..dc6705dd 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -10,6 +10,28 @@ The release workflow uses a tag's section here as its release notes (a
hand-written `docs/releases/<tag>.md` wins if present), so keeping this file
current is part of cutting a release.
+## v0.0.302.14
+
+### Models
+
+- MiMo thinks only at `high` effort and above (ADR 0235). graff's default effort is `medium`, which used to turn MiMo's thinking on for every call; MiMo then thought at length before each action, and that was most of a task's time. `minimal`, `low` and `medium` now send thinking Off; `/effort high`, a worker's effort pin or Jev's `high` turn it on. The picker, status line and ACP show the default as Off. (#1453)
+
+### Speed
+
+- MCP servers connect in the background in every session but `--json` (ADR 0230), so a slow server no longer holds the first message. Each request merges the servers that have finished connecting, `/mcp` names the ones still connecting, and ACP `session/new` waits at most 250 ms for a client's servers. (#1445)
+- Shell waits and `todo_write` replies stop costing turns (ADR 0233). Interactive sessions ignore a shorter shell `timeout`, so a quick command is no longer parked after a second and the turn no longer ends; unattended sessions keep a shorter wait but not below 5 s. `todo_write` replies with counts instead of repeating the list. (#1451)
+- The shell tool names the OS its commands run on, so models write BSD `sed -i ''` on macOS. The codedb guard (#626) lets in-place `sed` edits and pipe-free whole-file reads up to 16 KiB through; searches and bigger reads still go to codedb (ADR 0234). (#1452)
+- `rlm` takes JSON array and object literals as arguments and binds, `each()` maps plain values as well as objects, a lone object argument is the keyword arguments, and a refused statement lists the forms that work (ADR 0236). (#1454)
+
+### Sub-agents
+
+- Sub-agents at parity with the codex route (ADR 0232): `write_file` creates a missing directory, `agent_output` takes `ids`, a child starts with the user's prompts from its parent's history at the parent's live effort and can ask Jev for its own, read-only children on macOS run any command under a sandbox that refuses writes in the project and network access, and each finished child writes a `subagent` trace line. Unattended prompts leave out the `todo_write` nudges. (#1445)
+
+### Evals
+
+- `graff-evals` gains a `subagents` suite of five delegation tasks and its `subagents-solo` counterfactual, with `subagent_split.py` to split parent and child time. (#1445)
+- MiMo v2.6 Pro and Flash: v0.0.302.13 against this release and OpenCode 2, in `evals/mimo-v2.6/`. DeepSeek V4 Pro on DeepSeek's own API, graff against deepseek-harness, in `evals/graff-vs-deepseek-harness/`. `gpt-6-astra` beside the Codex app server, in `evals/graff-vs-codex-app-server/ASTRA.md`. See [docs/releases/v0.0.302.14.md](docs/releases/v0.0.302.14.md).
+
## v0.0.302.13
### Speed
diff --git a/docs/releases/v0.0.302.14.md b/docs/releases/v0.0.302.14.md
new file mode 100644
index 00000000..af97648f
--- /dev/null
+++ b/docs/releases/v0.0.302.14.md
@@ -0,0 +1,97 @@
+# Codegraff v0.0.302.14
+
+MiMo now answers without thinking unless you ask it to, MCP servers connect
+in the background, sub-agents work the way they do on the codex route, and
+shell waits and `todo_write` replies stop costing turns. Measured against the
+previous release and other harnesses on the same models:
+
+## Evals
+
+### MiMo v2.6 Pro: 2.9x as fast as v0.0.302.13
+
+The published v0.0.302.13 binary, this release and OpenCode 2 on the same
+model, endpoint and key, interleaved over 11 coding and 10 MCP tasks, 2 runs
+each:
+
+| | Pass | Wall per task | Coding tasks | MCP tasks | Output tokens per task |
+| --- | ---: | ---: | ---: | ---: | ---: |
+| v0.0.302.13 | 42/42 | 53.1s | 15.8s | 94.2s | 2,353 |
+| **v0.0.302.14** | 39/42 | **18.1s** | **11.5s** | **25.5s** | **544** |
+| OpenCode 2 | 42/42 | 30.9s | 17.4s | 45.7s | 1,004 |
+
+MiMo used to think at length before every action, and that was most of a
+task's time. Without thinking it missed 3 of 42 runs: one reply was a JSON
+fragment instead of a tool call, and one task twice reported the issues'
+display numbers instead of their ids. `/effort high` turns thinking back on
+for work that needs careful reading.
+
+MiMo v2.6 Flash, thinking on against this release's default: 35.1s against
+16.7s per task (OpenCode 2: 26.6s), 42/42 against 40/42.
+[Full results](https://github.com/justrach/codegraff/blob/v0.0.302.14/evals/mimo-v2.6/RESULTS.md).
+
+### DeepSeek V4 Pro: graff against deepseek-harness
+
+Both harnesses on DeepSeek's own API with one key, interleaved, 2 runs each:
+
+| | Pass | Wall per task | Coding tasks | MCP tasks | Model calls per MCP task |
+| --- | ---: | ---: | ---: | ---: | ---: |
+| **graff** | 42/42 | **22.6s** | **6.4s** | **40.4s** | **5.0** |
+| deepseek-harness 0.2.0-rc.2 | 42/42 | 27.7s | 7.4s | 50.2s | 9.2 |
+
+graff takes 19% less time over the 21 tasks. On the sub-agent suite it passed
+9 of 10 runs in 63.7s per task against 8 of 10 in 70.4s, and doing the same
+work without delegation 27.7s against 38.1s.
+[Full results](https://github.com/justrach/codegraff/blob/v0.0.302.14/evals/graff-vs-deepseek-harness/RESULTS.md).
+
+### gpt-6-astra: graff against the Codex app server, with sub-agents
+
+On the same model and ChatGPT account, graff is faster on every suite: 20% on
+the coding tasks, 30% on the MCP tasks, 18% when the tasks ask for sub-agents
+and 6% when they do not, at 16% to 73% lower cost on the app server's own
+reported usage.
+[Full results](https://github.com/justrach/codegraff/blob/v0.0.302.14/evals/graff-vs-codex-app-server/ASTRA.md).
+
+## MiMo thinks only at high effort
+
+graff's default effort is `medium`, and it used to turn MiMo's thinking on
+for every request. `minimal`, `low` and `medium` now send thinking Off on the
+Chat and Responses wires; `high` and above turn it on, and so do a worker's
+effort pin and Jev's `high`. The picker, status line and ACP show the default
+as Off ([ADR 0235](https://github.com/justrach/codegraff/blob/v0.0.302.14/docs/adr/0235-mimo-thinks-only-at-high-effort.md), #1453).
+
+## MCP connects in the background
+
+Every session but `--json` connects MCP servers in the background, so a slow
+server no longer holds the first message. Each request merges the servers
+that have finished, `/mcp` names the ones still connecting, and ACP
+`session/new` waits at most 250 ms for a client's servers
+([ADR 0230](https://github.com/justrach/codegraff/blob/v0.0.302.14/docs/adr/0230-mcp-connects-in-the-background.md), #1445).
+
+## Sub-agents at parity with the codex route
+
+- `write_file` creates a missing directory instead of refusing and making
+ the model regenerate the file.
+- `agent_output` takes `ids`, so collecting N children is one call.
+- A child starts with the user's prompts from its parent's history, runs at
+ its parent's live effort, and can ask Jev for its own.
+- Read-only children on macOS run any command under a sandbox that refuses
+ writes in the project and network access, so they can compute instead of
+ reporting partial results.
+- Each finished child writes a `subagent` trace line, and `graff-evals` has a
+ `subagents` suite with its `subagents-solo` counterfactual
+ ([ADR 0232](https://github.com/justrach/codegraff/blob/v0.0.302.14/docs/adr/0232-sub-agents-at-parity-with-the-codex-route.md), #1445).
+
+## Fewer wasted calls
+
+- Interactive sessions ignore a shorter shell `timeout`, so a quick command is
+ no longer parked after a second and the turn no longer ends. `todo_write`
+ replies with counts instead of repeating the list
+ ([ADR 0233](https://github.com/justrach/codegraff/blob/v0.0.302.14/docs/adr/0233-tool-results-the-model-already-has.md), #1451).
+- The shell tool names the OS, so models write BSD `sed -i ''` on macOS. The
+ codedb guard lets in-place `sed` edits and small whole-file reads through;
+ searches and big reads still go to codedb
+ ([ADR 0234](https://github.com/justrach/codegraff/blob/v0.0.302.14/docs/adr/0234-the-model-knows-the-os-and-the-codedb-guard-only-redirects-searches.md), #1452).
+- `rlm` takes JSON literals and binds them, `each()` maps plain values, a
+ lone object argument is the keyword arguments, and a refused statement lists
+ the forms that work
+ ([ADR 0236](https://github.com/justrach/codegraff/blob/v0.0.302.14/docs/adr/0236-rlm-takes-json-literals-and-each-maps-plain-values.md), #1454).
diff --git a/scripts/eval/tier1-manifest.json b/scripts/eval/tier1-manifest.json
index 9333e6cf..56ed229a 100644
--- a/scripts/eval/tier1-manifest.json
+++ b/scripts/eval/tier1-manifest.json
@@ -15,7 +15,7 @@
"src/repl.zig",
"TUI/root.zig"
],
- "test_count_baseline": 3131,
+ "test_count_baseline": 3154,
"test_count_slack": 25,
"required_invariants": [
{
diff --git a/src/cli.zig b/src/cli.zig
index 64d2aee8..d0b373da 100644
--- a/src/cli.zig
+++ b/src/cli.zig
@@ -23,6 +23,13 @@ const update_target = @import("update_target.zig");
pub const changelog_text =
\\What's new
\\──────────
+ \\0.0.302.14
+ \\ • MiMo answers without thinking unless effort is high (/effort high)
+ \\ • MCP servers connect in the background; the first message never waits
+ \\ • sub-agents start with your prompts and their parent's effort
+ \\ • the shell tool names the OS; small reads and sed -i pass the guard
+ \\ • rlm takes JSON literals; todo_write replies with counts
+ \\
\\0.0.302.13
\\ • fewer round trips: named MCP servers and subagents load up front
\\ • write_file won't replace an unread file and says what it did