agent-devkit 0.5.7 → 0.6.0
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 +27 -0
- package/dist/main.js +1293 -113
- package/dist/main.js.map +1 -1
- package/package.json +1 -1
- package/src/assets/i18n/en-US.json +16 -0
- package/src/assets/i18n/fr-FR.json +16 -0
- package/src/assets/i18n/ja-JP.json +16 -0
- package/src/assets/i18n/pt-BR.json +16 -0
- package/src/assets/i18n/zh-CN.json +16 -0
- package/src/assets/prompts/core/domain-decision.json +33 -0
package/README.md
CHANGED
|
@@ -200,6 +200,15 @@ agent sessions
|
|
|
200
200
|
agent sessions create --title "Planning"
|
|
201
201
|
agent sessions search "memory"
|
|
202
202
|
agent sessions resume <session-id>
|
|
203
|
+
agent runs
|
|
204
|
+
agent runs show <run-id>
|
|
205
|
+
agent runs events <run-id>
|
|
206
|
+
agent runs resume <run-id>
|
|
207
|
+
agent goals
|
|
208
|
+
agent goals create "Ship the release"
|
|
209
|
+
agent memory
|
|
210
|
+
agent memory learn "User prefers concise replies" --kind preference
|
|
211
|
+
agent integrations
|
|
203
212
|
agent tools
|
|
204
213
|
agent tools --json
|
|
205
214
|
agent install node --dry-run
|
|
@@ -255,6 +264,24 @@ Risky capabilities return `approval_required` unless `--approve` is passed. MCP
|
|
|
255
264
|
uses the same runtime contract, so future TUI and agent-loop execution should
|
|
256
265
|
also plug into this layer instead of calling capabilities directly.
|
|
257
266
|
|
|
267
|
+
## Agent Runtime, Memory and Integrations
|
|
268
|
+
|
|
269
|
+
The agentic runtime persists every run under `~/.agent-devkit/data/runtime/`
|
|
270
|
+
(`run.json`, `events.jsonl`, `plan.json`) with goals in `data/runtime/goals/`.
|
|
271
|
+
Runs are inspectable and resumable through `agent runs` and `agent goals`.
|
|
272
|
+
|
|
273
|
+
Global memory lives in the `memory` module: records have kind
|
|
274
|
+
(`fact`, `preference`, `pattern`, `decision`, `episode`), status
|
|
275
|
+
(`candidate`, `active`, `rejected`, `forgotten`), confidence, evidence refs
|
|
276
|
+
and an audit trail. Inferred learning enters as `candidate`; explicit user
|
|
277
|
+
input can enter as `active`. Chat and the task loop retrieve relevant
|
|
278
|
+
memory before responding.
|
|
279
|
+
|
|
280
|
+
External MCP servers are registered through `agent integrations` and their
|
|
281
|
+
tools are exposed to the agent through the composite tool runtime as
|
|
282
|
+
`mcp.<server>.<tool>`, classified by risk and requiring explicit approval
|
|
283
|
+
by default.
|
|
284
|
+
|
|
258
285
|
## MCP Server
|
|
259
286
|
|
|
260
287
|
Agent DevKit exposes its runtime tools through a real MCP server.
|