skillscript-runtime 0.14.0 → 0.14.1

package/examples/connectors/DataStoreTemplate/README.md

@@ -2,7 +2,7 @@
 
 A skeleton `DataStore` implementation for adopters writing their own. Not runnable; every method throws a `TODO` error. Copy this directory, rename, fill in the substrate-specific work.
 
-Use this when you want skillscript memories backed by:
+Use this when you want skillscript data persistence backed by:
 - A vector database (Pinecone, Weaviate, Qdrant, Chroma)
 - A data store backing (memory broker like AMP, hosted memory API, vector DB, etc.)
 - A different SQL flavor (Postgres + pgvector, MySQL)
@@ -18,7 +18,7 @@ DataStore (choose which connector)
    └── Your fork from this template
 ```
 
-The runtime is substrate-agnostic. Memories don't know which backend they're stored against — the contract is the `DataStore` interface, and any class implementing it works.
+The runtime is substrate-agnostic. Data records don't know which backend they're stored against — the contract is the `DataStore` interface, and any class implementing it works.
 
 ## Forking workflow
 
@@ -72,8 +72,8 @@ The DataStore contract is narrower than SkillStore. Three methods + `staticCapab
 
 | Method | What it does | When called |
 |---|---|---|
-| `query(filters)` | Read memories by mode + filter; return `PortableData[]` | Every `$ data_read mode=... query=...` op |
-| `write(entry)` | Persist a new memory; return `{id, created_at}` | Every `$ data_write content=... -> R` op + `data_write` skill notify routes |
+| `query(filters)` | Read records by mode + filter; return `PortableData[]` | Every `$ data_read mode=... query=...` op |
+| `write(entry)` | Persist a new record; return `{id, created_at}` | Every `$ data_write content=... -> R` op + `data_write` skill notify routes |
 | `manifest()` | Capability snapshot for `runtime_capabilities` discovery | At startup + on-demand from MCP clients |
 
 Per the curated-subset framing in `src/connectors/types.ts`, `PortableData` has a 4-tier field model:
@@ -113,7 +113,7 @@ Honor what your substrate supports; ignore the rest. Document the supported set
 
 The `DataWrite` shape:
 
-- **`content`** (required) — the memory body
+- **`content`** (required) — the record body
 - **`tags`** (optional) — routed to your substrate's tag mechanism
 - **`recipients`** (optional) — advisory hint; substrates with alerting (e.g., AMP's mailbox model) use it
 - **`expires_at`** (optional) — unix seconds; substrates with TTL honor it
@@ -121,7 +121,7 @@ The `DataWrite` shape:
 
 Return `{id, created_at}` — the substrate-assigned identifier + creation timestamp (unix seconds).
 
-If your substrate doesn't support writes (read-only memory like a search index over a static corpus), set `supports_writes: false` in `staticCapabilities()` and throw from `write()`. The runtime + lint will respect the flag.
+If your substrate doesn't support writes (read-only store like a search index over a static corpus), set `supports_writes: false` in `staticCapabilities()` and throw from `write()`. The runtime + lint will respect the flag.
 
 ## Wiring against the dashboard / MCP
 

package/examples/connectors/McpConnectorTemplate/README.md

@@ -139,7 +139,7 @@ The declarative path is the canonical adopter pattern for stdio-bridged remote M
 | Cardinality | Many instances per deployment | One singleton per slot |
 | Substrate config | Per-instance via top-level keys | `substrate` section short/object/custom |
 | Class extensibility | `registerConnectorClass()` for adopter-custom classes | Programmatic bootstrap (or `substrate.skill_store: {type: "custom", ...}` once async-bootstrap lands) |
-| Auto-wired bridges | `llm` + `memory` + `data_write` (LocalModel + DataStore exposed via bridge connectors) | n/a (these ARE the substrates being bridged) |
+| Auto-wired bridges | `llm` + `data_read` + `data_write` (LocalModel + DataStore exposed via bridge connectors) | n/a (these ARE the substrates being bridged) |
 
 McpConnector is fundamentally the "dispatch to external tools" surface — narrowest contract, broadest range of impls.
 

package/examples/custom-bootstrap.example.ts

@@ -86,14 +86,21 @@ registry.registerSkillStore("primary", skillStore);
 // Wire your own LocalModel, DataStore, AgentConnector here. For typed-
 // contract impls, this is the "case 1" wiring — substrate-portable.
 //   registry.registerLocalModel("default", new MyHostedLlmAdapter(...));
-//   registry.registerDataStore("primary", new MyMemoryAdapter(...));
+//   registry.registerDataStore("primary", new MyDataStoreAdapter(...));
 //   registry.registerAgentConnector("primary", new MyAgentHarnessAdapter(...));
 
-// Then wire bridges if you want the canonical `$ llm` / `$ data_read` surfaces:
+// Then wire bridges if you want the canonical `$ llm` / `$ data_read` /
+// `$ data_write` surfaces. The DataStore bridge handles BOTH read and
+// write via toolName discrimination; register the same instance under both
+// connector names so bare-form name-match resolution picks the right route.
 //   import { LocalModelMcpConnector, DataStoreMcpConnector } from "skillscript-runtime";
 //   registry.registerMcpConnector("llm", new LocalModelMcpConnector(registry.getLocalModel("default")));
 //   const ms = registry.listDataStores().find((e) => e.name === "primary");
-//   if (ms !== undefined) registry.registerMcpConnector("data_read", new DataStoreMcpConnector(ms.instance));
+//   if (ms !== undefined) {
+//     const dataBridge = new DataStoreMcpConnector(ms.instance);
+//     registry.registerMcpConnector("data_read", dataBridge);
+//     registry.registerMcpConnector("data_write", dataBridge);
+//   }
 
 // Wire any connectors.json instances.
 for (const c of connectors) {

package/examples/onboarding-scaffold/README.md

@@ -1,4 +1,4 @@
-# Onboarding scaffold — file-backed memory + OpenAI + tmux-shell
+# Onboarding scaffold — file-backed data store + OpenAI + tmux-shell
 
 A complete adopter-deployment example demonstrating substrate-portable wiring for skillscript-runtime. **~200 LOC across three adapter files** plus bootstrap. Copy this directory and modify for your own deployment.
 
@@ -25,7 +25,7 @@ cd ~/my-skillscript-deployment
 # 3. Set up env
 export SKILLSCRIPT_HOME=$(pwd)
 export OPENAI_API_KEY=sk-...
-cp memory.example.json memory.json  # initial memory
+cp data.example.json data.json  # initial data store seed
 
 # 4. Init skillscript dir layout
 skillfile init --here
@@ -49,7 +49,7 @@ node --loader ts-node/esm bootstrap.ts
 | `tmux-shell-agent-connector.ts` | `AgentConnector` impl — tmux send-keys | ~75 |
 | `bootstrap.ts` | Wiring — Registry, bridges, scheduler, MCP server | ~75 |
 | `connectors.json` | Example adopter-MCP wiring (empty by default) | — |
-| `memory.example.json` | Seed memory file with three example records | — |
+| `data.example.json` | Seed data file with three example records | — |
 
 ## Two-instance posture
 
@@ -77,4 +77,3 @@ The skill bodies don't need to change. `$ llm prompt=...` keeps working; `$ data
 
 - **Adopter playbook** — `docs/adopter-playbook.md` walks through Case 1 vs Case 2 wiring patterns
 - **Custom bootstrap walkthrough** — `examples/custom-bootstrap.example.ts` shows registering custom McpConnector classes via `registerConnectorClass`
-- **v0.8.x roadmap** — `$ data_write` ships in v0.8.x bundled with the auth model. When that lands, extend `FileDataStore` with the corresponding `write()` method.

package/examples/onboarding-scaffold/bootstrap.ts

@@ -1,4 +1,4 @@
-// Onboarding scaffold: complete bootstrap wiring file-backed memory +
+// Onboarding scaffold: complete bootstrap wiring file-backed data store +
 // OpenAI LLM + tmux-shell AgentConnector. v0.7.3.
 //
 // Copy this file into your deployment and modify substrate choices to
@@ -43,9 +43,9 @@ const registry = new Registry();
 const skillStore = new FilesystemSkillStore(config.skillsDir ?? `${HOME}/skills`);
 registry.registerSkillStore("primary", skillStore);
 
-// File-backed memory at $SKILLSCRIPT_HOME/memory.json
+// File-backed data store at $SKILLSCRIPT_HOME/data.json
 const dataStore = new FileDataStore({
-  filePath: config.dataDbPath ?? `${HOME}/memory.json`,
+  filePath: config.dataDbPath ?? `${HOME}/data.json`,
 });
 registry.registerDataStore("primary", dataStore);
 
@@ -63,10 +63,15 @@ const agentConnector = new TmuxShellAgentConnector({
 });
 registry.registerAgentConnector("primary", agentConnector);
 
-// Step 3: wire bridges so `$ llm` / `$ data_read` dispatch through the
-// adopter substrates above (case 1 typed-contract wiring — portable).
+// Step 3: wire bridges so `$ llm` / `$ data_read` / `$ data_write` dispatch
+// through the adopter substrates above (case 1 typed-contract wiring —
+// portable). The DataStore bridge handles BOTH read and write via toolName
+// discrimination; register the same instance under both connector names so
+// bare-form name-match resolution picks the right route.
 registry.registerMcpConnector("llm", new LocalModelMcpConnector(openai));
-registry.registerMcpConnector("data_read", new DataStoreMcpConnector(dataStore));
+const dataBridge = new DataStoreMcpConnector(dataStore);
+registry.registerMcpConnector("data_read", dataBridge);
+registry.registerMcpConnector("data_write", dataBridge);
 
 // Step 4: wire connectors.json instances (adopter-defined MCP servers).
 for (const c of connectors) {

package/examples/onboarding-scaffold/connectors.json

@@ -1,5 +1,5 @@
 {
-  "_comment": "Onboarding scaffold connectors.json — wires example MCP servers an adopter might add alongside the bundled bridges. Most onboarding deployments don't need anything here; the file-backed memory + OpenAI LLM + tmux-shell wiring lives in bootstrap.ts. This file is the place for adopter-specific MCP servers (your in-house tools, a YouTrack instance, a custom agent API, etc.).",
+  "_comment": "Onboarding scaffold connectors.json — wires example MCP servers an adopter might add alongside the bundled bridges. Most onboarding deployments don't need anything here; the file-backed data store + OpenAI LLM + tmux-shell wiring lives in bootstrap.ts. This file is the place for adopter-specific MCP servers (your in-house tools, a YouTrack instance, a custom agent API, etc.).",
 
   "_example_remote_mcp": {
     "class": "RemoteMcpConnector",

package/examples/onboarding-scaffold/file-data-store.ts

@@ -22,11 +22,11 @@ import type {
 } from "skillscript-runtime/connectors";
 
 export interface FileDataStoreConfig {
-  /** Absolute path to the JSON file holding the memory array. */
+  /** Absolute path to the JSON file holding the data records. */
   filePath: string;
 }
 
-interface FileMemoryRecord extends PortableData {
+interface FileDataRecord extends PortableData {
   /** Optional substrate-specific fields go in `metadata`; everything top-level matches PortableData. */
 }
 
@@ -77,7 +77,7 @@ export class FileDataStore implements DataStore {
     const created_at = Math.floor(Date.now() / 1000);
     const firstLine = entry.content.split("\n")[0] ?? entry.content;
     const summary = firstLine.length > 200 ? firstLine.slice(0, 197) + "..." : firstLine;
-    const newRecord: FileMemoryRecord = {
+    const newRecord: FileDataRecord = {
       id,
       summary,
       detail: entry.content,
@@ -92,7 +92,7 @@ export class FileDataStore implements DataStore {
             },
           }
         : {}),
-    } as FileMemoryRecord;
+    } as FileDataRecord;
     records.push(newRecord);
     writeFileSync(this.config.filePath, JSON.stringify(records, null, 2), "utf8");
     return { id, created_at };
@@ -106,19 +106,30 @@ export class FileDataStore implements DataStore {
         file_path: this.config.filePath,
         record_count: this.loadFile().length,
         supports_write: true,
+        // v0.14.1 strict-filters: the bridge enforces every non-base filter
+        // key in `query()` calls against this declared set, throwing
+        // UnsupportedFilterError for unknowns. This reference impl's
+        // substring scorer doesn't actually filter on any field (it
+        // ignores everything beyond `query` + `limit`), so the honest
+        // declaration is `[]`. Adopters who extend `query()` to honor
+        // `domain_tags` / `payload_type` / etc. should add those names
+        // here so the bridge stops rejecting them. Callers that need
+        // to pass advisory filters this impl ignores can opt out per-
+        // call via `permissive_filters: true`.
+        supported_filters: [],
       },
     };
   }
 
-  private loadFile(): FileMemoryRecord[] {
+  private loadFile(): FileDataRecord[] {
     if (!existsSync(this.config.filePath)) return [];
     try {
       const raw = readFileSync(this.config.filePath, "utf8");
       const parsed: unknown = JSON.parse(raw);
       if (!Array.isArray(parsed)) {
-        throw new Error(`FileDataStore: '${this.config.filePath}' top-level must be an array of memory records.`);
+        throw new Error(`FileDataStore: '${this.config.filePath}' top-level must be an array of data records.`);
       }
-      return parsed as FileMemoryRecord[];
+      return parsed as FileDataRecord[];
     } catch (err) {
       throw new Error(`FileDataStore: failed to read '${this.config.filePath}': ${(err as Error).message}`);
     }

package/examples/programmatic-trace-demo.mjs

@@ -34,7 +34,7 @@ const SKILL_SRC = `# Skill: heartbeat
 # Triggers: cron: */1 * * * *
 
 emit:
-    ! heartbeat at $(EVENT.fired_at_unix)
+    emit(text="heartbeat at $(EVENT.fired_at_unix)")
 
 default: emit
 `;

package/examples/skillscripts/classify-support-ticket.skill.md

@@ -1,30 +1,31 @@
 # Skill: classify-support-ticket
-# Status: Approved v1:99819a42
+# Status: Approved v1:670a94cf
+# Autonomous: true
 # Description: Read an incoming support ticket and route it: severity-1 tickets get paged to ops-channel, billing tickets get tagged for finance review, everything else gets a draft reply queued for human review
 # Vars: TICKET_TEXT, TICKET_ID
 # Output: agent: ops-oncall
 
 classify:
-    $ llm prompt="Categorize this support ticket. Reply with EXACTLY one of: 'sev-1', 'billing', 'general'. No other text.\n\nTicket: ${TICKET_TEXT}" model=qwen maxTokens=10 -> CATEGORY
+    $ llm prompt="Categorize this support ticket. Reply with EXACTLY one of: 'sev-1', 'billing', 'general'. No other text.\n\nTicket: ${TICKET_TEXT}" maxTokens=10 -> CATEGORY
 
 severity_check:
     needs: classify
-    $ llm prompt="Does this support ticket describe a service outage, data loss, security incident, or other production-severity-1 issue? Reply ONLY 'yes' or 'no'.\n\nTicket: ${TICKET_TEXT}" model=qwen maxTokens=10 -> IS_SEV1
+    $ llm prompt="Does this support ticket describe a service outage, data loss, security incident, or other production-severity-1 issue? Reply ONLY 'yes' or 'no'.\n\nTicket: ${TICKET_TEXT}" maxTokens=10 -> IS_SEV1
 
 route:
     needs: severity_check
     if ${CATEGORY|trim} == "sev-1":
         emit(text="PAGE: sev-1 ticket ${TICKET_ID} - ${TICKET_TEXT}")
-        $ datastore.write summary="sev-1 ticket ${TICKET_ID}" detail="${TICKET_TEXT}" knowledge_type=hard_won vault=private domain_tags=["support","sev-1","page"]
+        $ data_write content="sev-1 ticket ${TICKET_ID}: ${TICKET_TEXT}" tags=["support","sev-1","page"]
     elif ${IS_SEV1|trim} == "yes":
         emit(text="PAGE: classifier said ${CATEGORY|trim} but severity-check flagged this as sev-1: ${TICKET_ID}")
-        $ datastore.write summary="sev-1 escalation ${TICKET_ID}" detail="category=${CATEGORY|trim} but severity-check=yes" knowledge_type=hard_won vault=private domain_tags=["support","sev-1","disagreement"]
+        $ data_write content="sev-1 escalation ${TICKET_ID}: category=${CATEGORY|trim} but severity-check=yes" tags=["support","sev-1","disagreement"]
     elif ${CATEGORY|trim} == "billing":
         $set TAG_FOR = "finance"
         emit(text="Tagged for ${TAG_FOR} review: ${TICKET_ID}")
-        $ datastore.write summary="billing ticket ${TICKET_ID}" detail="${TICKET_TEXT}" knowledge_type=common vault=private domain_tags=["support","billing"]
+        $ data_write content="billing ticket ${TICKET_ID}: ${TICKET_TEXT}" tags=["support","billing"]
     else:
-        $ llm prompt="Draft a polite acknowledgment reply for this support ticket. Two short sentences. No greeting, no sign-off.\n\n${TICKET_TEXT}" model=qwen maxTokens=150 -> DRAFT
-        $ datastore.write summary="support draft ${TICKET_ID}" detail="${DRAFT|trim}" knowledge_type=common vault=private domain_tags=["support","draft"]
+        $ llm prompt="Draft a polite acknowledgment reply for this support ticket. Two short sentences. No greeting, no sign-off.\n\n${TICKET_TEXT}" maxTokens=150 -> DRAFT
+        $ data_write content="support draft ${TICKET_ID}: ${DRAFT|trim}" tags=["support","draft"]
 
 default: route

package/examples/skillscripts/cut-release-tag.skill.md

@@ -1,5 +1,5 @@
 # Skill: cut-release-tag
-# Status: Approved v1:eb29a798
+# Status: Approved v1:d3c2a91b
 # Description: Run when the user says "ship a release" or "cut a tag" on the current branch; collects commit-list since last tag, asks the user to confirm the version bump, applies the tag, and pushes it
 # Vars: BUMP_KIND=patch
 # Output: text
@@ -12,13 +12,13 @@ else:
 
 commits_since:
     needs: last_tag
-    shell(command="git log --oneline ${PREV_TAG}..HEAD") -> COMMITS
+    shell(command="git log --oneline \"${PREV_TAG}\"..HEAD") -> COMMITS
 else:
     $set COMMITS = "(unable to read commit log; proceeding without preview)"
 
 propose:
     needs: commits_since
-    $ llm prompt="Propose the next semver tag given prev tag '${PREV_TAG}', bump kind '${BUMP_KIND}', and commits below. Return ONLY the new tag string (e.g. v1.4.3). No commentary.\n\nCommits:\n${COMMITS}" model=qwen maxTokens=30 -> PROPOSED_TAG
+    $ llm prompt="Propose the next semver tag given prev tag '${PREV_TAG}', bump kind '${BUMP_KIND}', and commits below. Return ONLY the new tag string (e.g. v1.4.3). No commentary.\n\nCommits:\n${COMMITS}" maxTokens=30 -> PROPOSED_TAG
 
 confirm:
     needs: propose
@@ -30,11 +30,11 @@ confirm:
 
 apply:
     needs: confirm
-    shell(command="git tag ${PROPOSED_TAG|trim}") -> TAG_OUT
-    shell(command="git push origin ${PROPOSED_TAG|trim}") -> PUSH_OUT
+    shell(command="git tag \"${PROPOSED_TAG|trim}\"") -> TAG_OUT
+    shell(command="git push origin \"${PROPOSED_TAG|trim}\"") -> PUSH_OUT
     emit(text="Tagged and pushed ${PROPOSED_TAG|trim}")
 else:
     emit(text="Tag/push failed; rolling back local tag")
-    shell(command="git tag -d ${PROPOSED_TAG|trim}") -> ROLLBACK
+    shell(command="git tag -d \"${PROPOSED_TAG|trim}\"") -> ROLLBACK
 
 default: apply

package/examples/skillscripts/doc-qa-with-citations.skill.md

@@ -1,12 +1,12 @@
 # Skill: doc-qa-with-citations
 # Status: Approved v1:2187bf35
-# Description: When the user asks a question that requires retrieval over the doc set, answer with inline citations to memory IDs
+# Description: When the user asks a question that requires retrieval over the doc set, answer with inline citations to record IDs
 # Vars: QUESTION, K=6
 # Output: text
 
 answer:
     $ data_read mode=rerank query="${QUESTION}" limit=${K} -> HITS (fallback: [])
-    $ llm prompt="Answer the question using ONLY the supplied passages. Cite each claim inline as [id:<memory-id>]. Question: ${QUESTION}. Passages: ${HITS|json}" maxTokens=900 -> RESPONSE
+    $ llm prompt="Answer the question using ONLY the supplied passages. Cite each claim inline as [id:<record-id>]. Question: ${QUESTION}. Passages: ${HITS|json}" maxTokens=900 -> RESPONSE
     emit(text="${RESPONSE}")
 
 default: answer

package/examples/skillscripts/feedback-sentiment-scan.skill.md

@@ -1,12 +1,13 @@
 # Skill: feedback-sentiment-scan
-# Status: Approved v1:111610fc
+# Status: Approved v1:4484f7e0
+# Autonomous: true
 # Description: Each night, scan the previous 24h of customer feedback records, classify sentiment via local model, surface entries where sentiment is "frustrated" or "blocking" so the team sees them at start-of-day; skip entries already seen on prior nights
 # Triggers: cron: 0 3 * * *
 # Vars: SCAN_LIMIT=50
 # Output: agent: support-lead
 
 fetch_new:
-    $ data_read mode=fts query="customer feedback" limit=${SCAN_LIMIT} created_after=${EVENT.fired_at_unix} -> FEEDBACK
+    $ data_read mode=fts query="customer feedback" limit=${SCAN_LIMIT} -> FEEDBACK
 
 fetch_seen:
     $ data_read mode=fts query="sentiment-scan seen marker" limit=200 domain_tags=["sentiment-scan-seen"] -> SEEN_MARKERS
@@ -14,15 +15,15 @@ fetch_seen:
 classify_and_emit:
     needs: fetch_new, fetch_seen
     emit(text="Sentiment scan results for ${NOW}:")
-    foreach F in ${FEEDBACK}:
-        if ${F.id|trim} in ${SEEN_MARKERS}:
+    foreach F in ${FEEDBACK.items}:
+        if ${F.id|trim} in ${SEEN_MARKERS.items}:
             emit(text="- skipped (already classified): ${F.id|trim}")
-        elif ${F.id|trim} not in ${SEEN_MARKERS}:
-            $ llm prompt="Classify the sentiment of this customer feedback. Respond with ONE word: 'frustrated', 'blocking', 'satisfied', 'neutral'. No explanation.\n\nFeedback: ${F.summary}\nDetail: ${F.detail}" model=gemma2 maxTokens=10 -> VERDICT
+        elif ${F.id|trim} not in ${SEEN_MARKERS.items}:
+            $ llm prompt="Classify the sentiment of this customer feedback. Respond with ONE word: 'frustrated', 'blocking', 'satisfied', 'neutral'. No explanation.\n\nFeedback: ${F.summary}\nDetail: ${F.detail}" maxTokens=10 -> VERDICT
             if ${VERDICT|trim} == "frustrated":
                 emit(text="- FRUSTRATED [${F.id|trim}] ${F.summary}")
             elif ${VERDICT|trim} == "blocking":
                 emit(text="- BLOCKING [${F.id|trim}] ${F.summary}")
-            $ datastore.write summary="sentiment-scan seen ${F.id|trim}" detail="verdict=${VERDICT|trim} on ${EVENT.fired_at_unix}" knowledge_type=common vault=private domain_tags=["sentiment-scan-seen"] expires_at=${EVENT.fired_at_plus_7d_unix}
+            $ data_write content="sentiment-scan seen ${F.id|trim} verdict=${VERDICT|trim} on ${EVENT.fired_at_unix}" tags=["sentiment-scan-seen"] expires_at=${EVENT.fired_at_plus_7d_unix} -> ACK
 
 default: classify_and_emit

package/examples/skillscripts/hello.skill.provenance.json

@@ -2,7 +2,7 @@
   "provenance_version": "1.0",
   "language_version": "1.0",
   "compiler_version": "0.1.0-dev",
-  "compiled_at": "2026-05-30T14:28:14.260Z",
+  "compiled_at": "2026-05-31T19:21:57.734Z",
   "source_skill": {
     "name": "hello"
   },

package/examples/skillscripts/service-health-watch.skill.md

@@ -1,6 +1,7 @@
 # Skill: service-health-watch
-# Status: Approved v1:7c93b4b4
-# Description: Every 5 minutes check named service endpoints — if latency or status degrades, write a signal memory and alert
+# Status: Approved v1:628c8c69
+# Autonomous: true
+# Description: Every 5 minutes check named service endpoints — if latency or status degrades, write a signal record and alert
 # Vars: SERVICES=[auth-api, ledger-api, search-api], LATENCY_BUDGET_MS=400
 # Triggers: cron: */5 * * * *
 # Output: none
@@ -10,8 +11,8 @@ probe:
         shell(command="curl -s -o /dev/null -w \"%{http_code} %{time_total}\" https://status.internal/${SVC|url}") -> RAW
         $ llm prompt="From the line '${RAW|trim}' (http_code time_seconds), and budget ${LATENCY_BUDGET_MS} ms, answer ok or degraded only." -> STATUS
         if ${STATUS|trim} == "degraded":
-            $ datastore.write summary="service degradation: ${SVC}" detail="probe at ${NOW}: ${RAW|trim}" domain_tags=[ops, service-health, degraded:${SVC}] vault=private knowledge_type=common expires_at=${EVENT.fired_at_plus_1d_unix} -> ACK
-            emit(text="${SVC} degraded — wrote signal")
+            $ data_write content="service degradation: ${SVC} at ${NOW}: ${RAW|trim}" tags=["ops","service-health","degraded:${SVC}"] expires_at=${EVENT.fired_at_plus_1d_unix} -> ACK
+            emit(text="${SVC} degraded — wrote signal ${ACK.id}")
         else:
             emit(text="${SVC} ok")
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "skillscript-runtime",
-  "version": "0.14.0",
+  "version": "0.14.1",
   "description": "Runtime, compiler, lint, CLI, and dashboard for Skillscript — a small declarative language for authoring agent workflows.",
   "license": "MIT",
   "author": "Scott Shwarts <scotts@pobox.com>",

package/scaffold/config.toml

@@ -26,6 +26,8 @@ resolution_order = ["filesystem", "packages"]
 # See docs/configuration.md.
 
 [runtime]
-# `@@` opt-in unsafe shell. Disabled by default. v1 spec gates this behind
-# explicit opt-in plus a tier-2 lint warning.
+# `shell(command="...", unsafe=true)` (legacy `@ unsafe <body>`) — opt-in
+# full-bash shell. Disabled by default. Spec gates this behind explicit
+# opt-in plus a tier-2 lint warning that fires every appearance. Set to
+# `true` only after auditing every unsafe-shell op in your skill corpus.
 enable_unsafe_shell = false