zelari-code 2.6.3 โ 2.7.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +14 -0
- package/package.json +2 -2
package/README.md
CHANGED
|
@@ -261,6 +261,9 @@ Disable auto-check: `ANATHEMA_DEV=1 zelari-code`
|
|
|
261
261
|
- ๐ค **Multi-agent council** โ 6 roles (Caronte, Nettuno, Gerione, Plutone, Minosse, Lucifero) with feedback loops and member promotion
|
|
262
262
|
- ๐ **Kraken super-agent** โ default mode (aliases `agent`/`single`): lead that spawns `task` tentacles (`explore` / `general` / `verify`), optional git worktrees, and **Kraken Graph** (`/kraken graph`) for a parallel DAG
|
|
263
263
|
- โก **Zelari-mode** โ autonomous multi-run missions: a free-form prompt becomes a mission brief, then **design@council โ build@kraken** until the MVP slice's `completion.ok` is green or the iteration budget runs out
|
|
264
|
+
- ๐งฎ **Budget-aware mission continuation** โ after each mission slice a continuation gate decides repair / pivot / hold from the remaining budget and gap history (repeated gap โ pivot with reduced roster; exhaustion โ hold, never a false done; deterministic PASS stays the only authority)
|
|
265
|
+
- ๐งพ **Deep harness manifest** โ each session fingerprints its real tool surface (name + description + input schema per tool) alongside profile and resource policy, so tool or schema drift changes the manifest hash and is detectable on resume and in eval provenance
|
|
266
|
+
- ๐ฆ **Eval retention gate** โ tiered anchor suite (`eval/anchors/`), versioned result store per manifest hash, and a CI regression gate comparing candidate vs last stable tag (blocking once the baseline store is seeded; honest shadow + warning until then)
|
|
264
267
|
- ๐ง **Project memory** โ zero-dependency file-based recall (`.zelari/memory/`), fed into the council as RAG context between mission slices (opt-out with `ZELARI_MEMORY=0`)
|
|
265
268
|
- โงโฅ **Kraken/council/zelari mode switch** โ `shift+tab` cycles free-form prompts between the kraken lead, the full council pipeline, and an autonomous mission (mode shown in the status line)
|
|
266
269
|
- ๐จ **Rich TUI** โ Ink + React: native-scrollback chat stream, input bar with status line below it (mode ยท provider ยท model ยท session ยท cwd ยท execution timer)
|
|
@@ -338,6 +341,8 @@ zelari-code (CLI, Apache-2.0)
|
|
|
338
341
|
| `ZELARI_MEMORY=0` | Disable the file-based project memory (`.zelari/memory/`) |
|
|
339
342
|
| `ZELARI_MISSION_AUTO=1` | Auto-start Zelari missions (skip the brief confirmation) |
|
|
340
343
|
| `ZELARI_MISSION_MAX_ITER` | Max **implementation** slices per Zelari mission (default 6; design-phase is free; impl 2+ = Minosse+Lucifero only) |
|
|
344
|
+
| `ZELARI_TASK_CONTRACT=0` | Disable the mission TaskContract (goal / constraints / acceptance criteria extracted from the brief; on by default) |
|
|
345
|
+
| `ZELARI_EVAL_RESULTS_DIR` | Dev: override the eval result-store dir used by `tools/eval` (test/CI) |
|
|
341
346
|
| `ZELARI_MODE_MAX_TOOLS_LUCIFER` | Chairman (Lucifero) tool budget in zelari-mode (default 30) |
|
|
342
347
|
| `ZELARI_DIAGNOSTICS=0` | Disable the post-edit compiler/lint diagnostics loop |
|
|
343
348
|
| `ZELARI_DIAGNOSTICS_TIMEOUT_MS` | Timeout of the diagnostics loop (default 5000) |
|
|
@@ -458,6 +463,15 @@ npm test
|
|
|
458
463
|
npm run typecheck
|
|
459
464
|
```
|
|
460
465
|
|
|
466
|
+
### Eval retention gate
|
|
467
|
+
|
|
468
|
+
```bash
|
|
469
|
+
npm run test:eval # eval engine + regression-gate unit tests
|
|
470
|
+
npm run eval:gate # regression gate vs stored baseline (CI parity)
|
|
471
|
+
```
|
|
472
|
+
|
|
473
|
+
Anchors: `eval/anchors/` (tiered) ยท result store: `eval/results/`, versioned per harness-manifest hash and seeded with `node --experimental-strip-types tools/eval/runAnchors.ts` ยท CI: `.github/workflows/eval-retention-gate.yml` โ Tier-0 always blocking; the retention gate runs candidate anchors vs the store extracted from the last stable tag and **blocks on regression** once the baseline store is seeded (declared bootstrap + warning until then; never fake greens).
|
|
474
|
+
|
|
461
475
|
## Related
|
|
462
476
|
|
|
463
477
|
- [Product page](https://anathema-studio.com/zelari-code) (IT: [zelari-codice](https://anathema-studio.com/zelari-codice)) โ Anathema Studio
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "zelari-code",
|
|
3
|
-
"version": "2.
|
|
3
|
+
"version": "2.7.0",
|
|
4
4
|
"description": "Zelari Code โ AI Council coding agent CLI. Multi-agent orchestration (Caronte, Nettuno, Gerione, Plutone, Minosse, Lucifero) with slash commands, provider-agnostic LLM streaming, and self-update.",
|
|
5
5
|
"author": "Anathema Studio <https://anathema-studio.com/>",
|
|
6
6
|
"license": "Apache-2.0",
|
|
@@ -59,7 +59,7 @@
|
|
|
59
59
|
},
|
|
60
60
|
"devDependencies": {
|
|
61
61
|
"@testing-library/react": "^16.3.2",
|
|
62
|
-
"@zelari/core": "2.
|
|
62
|
+
"@zelari/core": "2.7.0",
|
|
63
63
|
"@types/node": "^25.3.0",
|
|
64
64
|
"@types/react": "^19.0.10",
|
|
65
65
|
"esbuild": "^0.25.0",
|