@tangle-network/agent-knowledge 1.9.0 → 1.10.0
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 +6 -0
- package/dist/index.d.ts +1380 -1080
- package/dist/index.js +764 -390
- package/dist/index.js.map +1 -1
- package/dist/memory/index.d.ts +1 -1
- package/dist/{types-BEDGlXB-.d.ts → types-C17sAROL.d.ts} +1 -1
- package/dist/viz/index.d.ts +1 -1
- package/docs/results/adaptive.md +120 -0
- package/docs/results/claim-grounding.md +97 -0
- package/docs/results/cost-quality.md +27 -0
- package/docs/supervisor-profiles.md +295 -0
- package/docs/two-agent-research-ab.md +428 -0
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -271,6 +271,12 @@ with its own pass (`driverResearches: true`), and gates on the readiness check.
|
|
|
271
271
|
Both are yours (no creds) — the loop owns the deterministic mechanics (indexing,
|
|
272
272
|
applying write blocks, scoring readiness) and stops once no blocking gap remains.
|
|
273
273
|
|
|
274
|
+
Does the verifying driver actually earn its keep? See
|
|
275
|
+
[docs/two-agent-research-ab.md](docs/two-agent-research-ab.md) for an equal-compute
|
|
276
|
+
A/B (9 ML topics, `glm-5.2`): the two-agent loop admits ~2.33 fewer sources per topic
|
|
277
|
+
at identical coverage — though most of that win is de-duplication, not relevance
|
|
278
|
+
filtering. Honest caveats and how to reproduce included.
|
|
279
|
+
|
|
274
280
|
```ts
|
|
275
281
|
import {
|
|
276
282
|
defineReadinessSpec,
|