@tangle-network/agent-runtime 0.168.0 → 0.170.0

package/dist/testing.js

@@ -1,12 +1,12 @@
 import { ct as immutableCandidateValue, et as parseExactAgentProfile, it as canonicalCandidateDigest$1, q as applyExactAgentProfileDiff } from "./prepare-BAyaNoZu.js";
-import { f as verifyAgentImprovementProposal } from "./improvement-cycle-BAKN_f4x.js";
+import { f as verifyAgentImprovementProposal } from "./improvement-cycle-BYCnKZ_G.js";
 import { c as supervisorAgentWithTestBrain, ft as driverAgent, r as superviseWithTestBrain } from "./supervise-rEE46ruH.js";
 import { i as runGraphWithTestBrain } from "./graph-ZDMygFIo.js";
 import { SANDBOX_SIZE_PRESET_NAMES } from "@tangle-network/agent-interface";
 //#region src/testing/fixtures/agent-improvement-proposal.json
 var agent_improvement_proposal_default = {
 	changedSurfaces: ["prompt"],
-	digest: "sha256:11d743eaab24db35601c84e2f17a7b36dbf2974fb5678838360293f220a969ef",
+	digest: "sha256:62b716af557ebbb5fb1ac82cb343bef5d5d9333afc4dffaa642ef03a8023b047",
 	evaluation: {
 		"decision": {
 			"contributingChecks": [
@@ -4577,7 +4577,7 @@ var agent_improvement_proposal_default = {
 		],
 		"metadata": {
 			"fixture": "agent-improvement-proposal",
-			"runtimeVersion": "0.168.0"
+			"runtimeVersion": "0.170.0"
 		},
 		"objectives": [
 			{
@@ -4688,8 +4688,8 @@ var agent_improvement_proposal_default = {
 			"baselineContentHash": "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
 			"candidateContentHash": "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
 			"kind": "agent-eval-loop",
-			"recordDigest": "sha256:aa5899c94f12c8551e49a34eed6a273e3eaa576edfc29f08fccae2f587b692b6",
-			"runId": "agent-runtime-0.168.0-proposal-fixture",
+			"recordDigest": "sha256:9d28a3e26e57d69dd9798413154f710161b43b635bf0b19f182aa87fb3c390af",
+			"runId": "agent-runtime-0.170.0-proposal-fixture",
 			"schema": "agent-candidate-experiment"
 		}
 	},
@@ -4712,13 +4712,13 @@ var agent_improvement_proposal_default = {
 	}],
 	kind: "agent-improvement-proposal",
 	proposedAt: "2026-07-10T01:00:00.000Z",
-	runId: "agent-runtime-0.168.0-proposal-fixture"
+	runId: "agent-runtime-0.170.0-proposal-fixture"
 };
 //#endregion
 //#region src/testing/fixtures/agent-profile-improvement-proposal.json
 var agent_profile_improvement_proposal_default = {
 	changedSurfaces: ["prompt", "skills"],
-	digest: "sha256:c869e70215e5f34bf9fc72edbfab994e0591db72eaab5bbc43fd2506867a2c1f",
+	digest: "sha256:6f1d98a819ac4bb1c873eb79ea0e430acee97f1dd11714dacb32b27b9f46748c",
 	evaluation: {
 		"decision": {
 			"contributingChecks": [
@@ -6352,7 +6352,7 @@ var agent_profile_improvement_proposal_default = {
 		],
 		"metadata": {
 			"fixture": "agent-profile-improvement-proposal",
-			"runtimeVersion": "0.168.0"
+			"runtimeVersion": "0.170.0"
 		},
 		"objectives": [
 			{
@@ -6463,7 +6463,7 @@ var agent_profile_improvement_proposal_default = {
 			"baselineContentHash": "sha256:21c495a37c418c10bde64fbaa188beddeed31f1f051ea60a6a6582a9ee0db704",
 			"candidateContentHash": "sha256:103f77bc8481601eef1ad5fe6ba84a40dffabc3a44f421f8c8559121edab84e9",
 			"kind": "agent-eval-loop",
-			"recordDigest": "sha256:3fd2e658bd7ce49643b936205aa5701f516f415d751f935ee05cb8e8ab30d0a2",
+			"recordDigest": "sha256:1e7e4b7cf942418b42ccb1eae9db4c237bce187aac1d8ac2fb63e76367e9e6c4",
 			"runId": "profile-improvement-1",
 			"schema": "agent-profile-improvement-experiment"
 		}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tangle-network/agent-runtime",
-  "version": "0.168.0",
+  "version": "0.170.0",
   "description": "Shared task-lifecycle skeleton for agents: a recursive loop kernel for chat turns, one-shot tasks, and multi-attempt loops, with trace capture and eval-gated self-improvement. Domain behavior lives in adapters; scoring and ship-gates in @tangle-network/agent-eval.",
   "homepage": "https://github.com/tangle-network/agent-runtime#readme",
   "repository": {

package/skills/codemode/SKILL.md

@@ -0,0 +1,48 @@
+---
+name: codemode
+description: Batch mechanical tool work as one program so loops and intermediates stay out of context.
+---
+
+# Codemode
+
+Use this policy when a task needs three or more mechanical tool or command calls whose intermediate results need no judgment.
+One call per model turn spends a round trip per step and pushes every intermediate value through the context window.
+Write one program instead: the loop, the branch, and the intermediates stay in the program, and only the decision-relevant summary returns.
+
+This is the pattern the ecosystem calls code mode (Cloudflare's Code Mode, Anthropic's code execution with MCP, the CodeAct paper).
+In a coding harness you already have the whole capability: a shell, a filesystem, and the tools this profile grants.
+
+## Run The Work
+
+1. List the calls the task needs and mark which results require your judgment.
+2. Put every judgment-free stretch into one script; keep each judgment point in your own turn.
+3. Hold intermediates in variables or files inside the workspace, never in your reply.
+4. Make the script print only the decision-relevant summary: counts, failures, the final value.
+5. Prefer one script that fans out over N items to N separate tool calls with identical shape.
+6. Stop batching the moment a result changes what you would do next; read it, decide, then batch again.
+
+## Boundaries That Are Not Yours To Move
+
+Spawning, steering, and settling agents go through your coordination tools, never through a script.
+A script that reaches those verbs over HTTP bypasses the budget pool and the journal, so the run's spend and record lie.
+An operation that costs money must run where the runtime meters it; do not wrap metered work in a script that hides the spend.
+The lint on authored code refuses imports, `process`, and network access; it is a lint, not a sandbox, so treat generated code you did not review as untrusted.
+
+## Router-Brained Agents
+
+A raw chat model has no shell, so this policy does not apply to it directly.
+Give such a node a code action space with the graph engine's `codemode` kind: an `operations` table projects the API the model sees, a host `codeRunner` executes, and each operation's spend reaches the settlement.
+A supervisor that only needs one small computation can carry an `extraTools` entry instead of a codemode node.
+
+## Common Mistakes
+
+- Batching a step whose output should have changed your plan, then discovering it three steps later.
+- Printing a whole dataset into the reply instead of writing it to a file and printing the summary.
+- Re-running an expensive script to re-read a value the first run already produced; write results to files.
+- Moving supervision into a script because the coordination verbs are reachable over local HTTP.
+
+## Then consider
+
+- `supervise` when the batched work is really delegation to workers with their own judgment.
+- `agent-graphs` when the shape of the work is a fixed topology rather than one agent's loop.
+- `loop-writer` when no shipped composition API can express the control policy you need.