@vpxa/aikit 0.1.249 → 0.1.251

package/packages/viewers/README.md

@@ -93,6 +93,7 @@ Important distinction:
 - Root `*-viewer.html` files are Vite entry templates
 - `dist/*.html` files are the final self-contained artifacts
 - `embed.mjs` reads from `dist/`, not from the root entry templates
+- `package.json` `exports` points at the built viewer HTML assets, which keeps workspace health satisfied without changing the standalone build flow
 
 ## Architecture
 

package/scaffold/dist/definitions/bodies.mjs

@@ -149,7 +149,7 @@ import{postTaskLesson as e,preTaskKnowledgeRecall as t}from"./protocols.mjs";con
   **Subagent prompt template:**
   1. **Scope** — exact files + boundary
   2. **Goal** — acceptance criteria, testable
-  3. **Arch Context** — pick by \`config.tokenBudget\`: efficient → \`stratum_card({ files: ['<path>'], query: '<what matters>', tier: 'T1' })\`, normal → \`compact({path, query})\`, full → \`digest({ sources: [...], query: '<what matters>' })\`. Default to efficient.
+  3. **Arch Context** — pick by \`config.tokenBudget\`: efficient → \`stratum_card({ files: ['<path>'], query: '<what matters>', tier: 'T1' })\`, normal → \`compact({path, query})\` or \`compact({ref, query?})\`, full → \`digest({ sources: [...], query: '<what matters>' })\`. Default to efficient.
   4. **Constraints** — patterns, conventions
   5. **Prior Knowledge** — Fetch topic-scoped knowledge: \`knowledge({ action: "lesson", subAction: "list-lessons", topic: "<2-3 task keywords>", minConfidence: 70 })\` + \`search({ query: "<task area>", category: "conventions", limit: 3 })\`. Include HIGH-confidence results (≥70) under \`## Prior Knowledge\`. Skip if none.
   6. **Artifacts Path** — the active flow's run directory and artifacts path from \`flow({ action: 'status' })\` (e.g. \`.flows/add-authentication/.spec/\`)
@@ -188,7 +188,7 @@ import{postTaskLesson as e,preTaskKnowledgeRecall as t}from"./protocols.mjs";con
 
   ### Context Gathering for Subagent Prompts
 
-  Default to \`stratum_card({ files: ['<path>'], query: '<what matters>', tier: 'T1' })\`; upgrade to \`compact\` or \`digest\`; use \`read_file\` only for exact edit lines.
+  Default to \`stratum_card({ files: ['<path>'], query: '<what matters>', tier: 'T1' })\`; upgrade to \`compact({ path, query })\`, \`compact({ ref, query? })\`, or \`digest\`; use \`read_file\` only for exact edit lines.
 
   **Knowledge injection (MANDATORY for Standard+ tier):** Before any subagent prompt, call:
   - \`knowledge({ action: "lesson", subAction: "list-lessons", topic: "<task keywords>", minConfidence: 70 })\`
@@ -415,7 +415,7 @@ import{postTaskLesson as e,preTaskKnowledgeRecall as t}from"./protocols.mjs";con
 ## Pre-Edit Checklist
 
 1. **Understand consumers** — \`graph({action:'find_nodes', name_pattern:'<target>'})\` → \`graph({action:'neighbors', node_id, direction:'incoming'})\`
-2. **Compress, don't raw-read** — \`file_summary\` then \`compact({path, query})\`; \`read_file\` only for exact edit lines
+2. **Compress, don't raw-read** — \`file_summary\` then \`compact({path, query})\` or \`compact({ref, query?})\`; \`read_file\` only for exact edit lines
 3. **Snapshot risky edits** — \`checkpoint({action:'save', label:'pre-<scope>'})\` before cross-cutting changes
 4. **Estimate blast radius** — run \`blast_radius\` before and after shared/public symbol changes
 5. **TDD when tests exist** — failing test first, then minimum code
@@ -530,6 +530,8 @@ ${e()}
 
 ### Phase 2: Reproduce
 
+Tag filters are exact labels, not substring search.
+
 1. \`search({ query: "<error-keywords>", tags: ["observation"] })\`
 2. \`search({ query: "error patterns" })\`
 3. \`knowledge({ action: "list", tag: "errors" })\`

package/scaffold/dist/definitions/flows.mjs

@@ -10,7 +10,7 @@ Follow \`multi-agents-development\` dispatch patterns:
 2. **Subagent Context**:
    - **Scope**: exact files + do NOT touch boundary
    - **Goal**: testable acceptance criteria
-   - **Arch Context**: code via \`compact()\`/\`digest()\`
+   - **Arch Context**: code via \`compact()\`/\`digest()\` or a cached ref via \`compact({ ref, query? })\`
    - **Constraints**: patterns, conventions, anti-patterns
    - **Self-Review**: checklist before DONE
 

package/scaffold/dist/definitions/protocols.mjs

@@ -52,6 +52,7 @@ Use AI Kit retrieval/compression first. Native tools are fallback only.
 | \`read_file\` to understand a file | \`file_summary({ path })\` | Structure first |
 | \`read_file\` to find code | \`compact({ path, query })\` | Focused extract |
 | Multiple \`read_file\` calls | \`digest({ sources, query: "<task description>" })\` | Compress multi-file context |
+| Cached ref to refocus prior output | \`compact({ ref, query? })\` | Reuse reversible text artifacts |
 | \`grep_search\` / \`semantic_search\` | \`search({ query })\` | Indexed search |
 | \`grep_search\` for a symbol | \`symbol({ name })\` | Def + refs |
 | \`run_in_terminal\` for tsc/lint | \`check({})\` | Narrow validation |
@@ -196,6 +197,7 @@ Use targeted searches before expensive work:
 | \`symbol({ name })\` | Definition + refs |
 | \`trace({ start, direction })\` | Call/data flow |
 | \`compact({ path, query })\` | Read specific section of a file |
+| \`compact({ ref, query? })\` | Reuse cached search/find/knowledge/compact output |
 | \`read_file\` | **ONLY** when you need exact lines for a pending edit |
 
 ### Step 4: Tool Discovery

package/scaffold/dist/definitions/skills/aikit.mjs

@@ -155,9 +155,9 @@ Every tool call passes through extractors that capture reusable facts:
 | **Research results** | \`web_search\`, \`web_fetch\`, \`search\`, \`find\` | Search findings, web content, code discoveries |
 | **Behavioral patterns** | ALL tools (ring buffer) | Tool preferences, search→edit sequences, error-fix cycles |
 
-**Retrieval:** Use \`enrich: true\` on any tool that supports it — the server automatically surfaces previously captured facts relevant to your query. Or search directly: \`search({ query: "...", origin: "curated" })\`
+**Retrieval:** Use \`enrich: true\` on any tool that supports it — the server automatically surfaces previously captured facts relevant to your query. Search/find/knowledge responses can also emit reusable refs; feed those into \`compact({ ref, query })\` instead of reopening the source.
 
-**Best practice:** Before re-reading a file you've already explored, try \`search({ query: "file-path-or-topic" })\` — auto-knowledge may already have what you need.
+**Best practice:** Before re-reading a file you've already explored, try \`search({ query: "file-path-or-topic" })\` — auto-knowledge may already have what you need, or reuse an existing \`ref\` with \`compact({ ref, query })\`.
 
 ### Retention & Tier Promotion
 - Memories start at **working** tier → promote on repeated access: episodic (2) → semantic (5) → procedural (10)

package/packages/store/dist/lance-store-BRKcJXVO.js

@@ -1 +0,0 @@
-import{EMBEDDING_DEFAULTS as e,SEARCH_DEFAULTS as t,STORE_DEFAULTS as n,createLogger as r,serializeError as i,sourceTypeContentTypes as a}from"../../core/dist/index.js";import{Index as o,connect as s}from"@lancedb/lancedb";function c(e){if(!e)return[];try{let t=JSON.parse(e);return Array.isArray(t)?t:[]}catch{return[]}}const l=/^[\w.\-/ ]+$/,u=r(`store`);function d(e,t){if(!l.test(e))throw Error(`Invalid ${t} filter value: contains disallowed characters`);return e.replace(/'/g,`''`)}var f=class r{db=null;table=null;dbPath;tableName;_draining=!1;_priorityQueue=[];_normalQueue=[];_ftsReady=!1;_ftsRecoveryAttemptAt=0;static FTS_RECOVERY_COOLDOWN_MS=300*1e3;enqueueWrite(e,t=!1){return new Promise((n,r)=>{let i=async()=>{try{n(await e())}catch(e){r(e)}};t?this._priorityQueue.push(i):this._normalQueue.push(i),this._drain()})}async _drain(){if(!this._draining){this._draining=!0;try{for(;this._priorityQueue.length>0||this._normalQueue.length>0;){let e=this._priorityQueue.shift()??this._normalQueue.shift();e&&await e()}}finally{this._draining=!1}}}constructor(e){this.dbPath=e?.path??n.path,this.tableName=e?.tableName??n.tableName}async initialize(){this.db=await s(this.dbPath),(await this.db.tableNames()).includes(this.tableName)&&(this.table=await this.db.openTable(this.tableName),await this.createFtsIndex())}async upsert(e,t){if(e.length!==0){if(e.length!==t.length)throw Error(`Record count (${e.length}) does not match vector count (${t.length})`);for(let n=0;n<t.length;n++)if(t[n].length===0)throw Error(`Zero-length vector at index ${n} for record ${e[n].sourcePath}`);return this.enqueueWrite(()=>this._upsertImpl(e,t))}}async upsertInteractive(e,t){if(e.length!==0){if(e.length!==t.length)throw Error(`Record count (${e.length}) does not match vector count (${t.length})`);for(let n=0;n<t.length;n++)if(t[n].length===0)throw Error(`Zero-length vector at index ${n} for record ${e[n].sourcePath}`);return this.enqueueWrite(()=>this._upsertImpl(e,t),!0)}}async _upsertImpl(e,t){let n=e.map((e,n)=>({id:e.id,vector:Array.from(t[n]),content:e.content,sourcePath:e.sourcePath,contentType:e.contentType,headingPath:e.headingPath??``,chunkIndex:e.chunkIndex,totalChunks:e.totalChunks,startLine:e.startLine,endLine:e.endLine,fileHash:e.fileHash,indexedAt:e.indexedAt,origin:e.origin,tags:JSON.stringify(e.tags),category:e.category??``,version:e.version}));if(this.table){let t=[...new Set(e.map(e=>e.sourcePath))];for(let e of t)try{await this.table.delete(`sourcePath = '${d(e,`sourcePath`)}'`)}catch{}await this.table.add(n)}else try{this.table=await this.db?.createTable(this.tableName,n)??null}catch(e){if(String(e).includes(`already exists`)&&this.db)this.table=await this.db.openTable(this.tableName),await this.table.add(n);else throw e}}async search(e,n){if(e.length===0||!this.table)return[];let r=n?.limit??t.maxResults,i=n?.minScore??t.minScore,a=this.table.search(e).limit(r*2),o=this.buildFilterString(n);return o&&(a=a.where(o)),(await a.toArray()).map(e=>({record:this.fromLanceRecord(e),score:1-(e._distance??1)})).filter(e=>e.score>=i).slice(0,r)}async createFtsIndex(){return this.enqueueWrite(()=>this._createFtsIndexImpl())}async _createFtsIndexImpl(){if(this.table)try{await this.table.createIndex(`content`,{config:o.fts({withPosition:!0}),replace:!0}),this._ftsReady=!0,this._ftsRecoveryAttemptAt=0,u.info(`FTS index created/updated`,{column:`content`})}catch(e){u.warn(`FTS index creation failed`,i(e))}}async ftsSearch(e,n){if(!this.table)return[];if(!this._ftsReady){let e=Date.now();if(e-this._ftsRecoveryAttemptAt<r.FTS_RECOVERY_COOLDOWN_MS)return[];this._ftsRecoveryAttemptAt=e;try{await this.createFtsIndex()}catch{return[]}if(!this._ftsReady)return[]}let a=n?.limit??t.maxResults;try{let t=this.table.search(e).limit(a*2),r=this.buildFilterString(n);return r&&(t=t.where(r)),(await t.toArray()).map(e=>({record:this.fromLanceRecord(e),score:e._score??e._relevance_score??0}))}catch(e){return(e instanceof Error?e.message:String(e)).includes(`INVERTED index`)?(u.debug(`FTS search skipped — index not yet available`),this._ftsReady=!1):u.warn(`FTS search failed`,i(e)),[]}}async getById(e){if(!this.table)return null;let t=await this.table.query().where(`id = '${d(e,`id`)}'`).limit(1).toArray();return t.length===0?null:this.fromLanceRecord(t[0])}async deleteBySourcePath(e){return this.enqueueWrite(()=>this._deleteBySourcePathImpl(e))}async _deleteBySourcePathImpl(e){if(!this.table)return 0;let t=await this.getBySourcePath(e);return t.length===0?0:(await this.table.delete(`sourcePath = '${d(e,`sourcePath`)}'`),t.length)}async deleteById(e){return this.enqueueWrite(()=>this._deleteByIdImpl(e))}async deleteByIdInteractive(e){return this.enqueueWrite(()=>this._deleteByIdImpl(e),!0)}async _deleteByIdImpl(e){return!this.table||!await this.getById(e)?!1:(await this.table.delete(`id = '${d(e,`id`)}'`),!0)}async getBySourcePath(e){return this.table?(await this.table.query().where(`sourcePath = '${d(e,`sourcePath`)}'`).limit(1e3).toArray()).map(e=>this.fromLanceRecord(e)):[]}async getStats(){if(!this.table)return{totalRecords:0,totalFiles:0,contentTypeBreakdown:{},lastIndexedAt:null,storeBackend:`lancedb`,embeddingModel:e.model};let t=await this.table.countRows(),n=await this.table.query().select([`sourcePath`,`contentType`,`indexedAt`]).limit(1e5).toArray(),r={},i=new Set,a=null;for(let e of n){let t=e;r[t.contentType]=(r[t.contentType]??0)+1,i.add(t.sourcePath),(!a||t.indexedAt>a)&&(a=t.indexedAt)}return{totalRecords:t,totalFiles:i.size,contentTypeBreakdown:r,lastIndexedAt:a,storeBackend:`lancedb`,embeddingModel:e.model}}async listSourcePaths(){if(!this.table)return[];let e=await this.table.query().select([`sourcePath`]).limit(1e5).toArray();return[...new Set(e.map(e=>e.sourcePath))]}async dropTable(){return this.enqueueWrite(()=>this._dropTableImpl())}async _dropTableImpl(){if(this.db&&(await this.db.tableNames()).includes(this.tableName))for(let e=1;e<=3;e++)try{await this.db.dropTable(this.tableName);break}catch(t){if(e===3)throw t;let n=e*500;u.warn(`dropTable attempt failed, retrying`,{attempt:e,delayMs:n}),await new Promise(e=>setTimeout(e,n))}this.table=null}async close(){try{this.db&&typeof this.db.close==`function`&&await this.db.close()}catch{}this.table=null,this.db=null}buildFilterString(e){let t=[];if(e?.contentType&&t.push(`contentType = '${d(e.contentType,`contentType`)}'`),e?.sourceType){let n=a(e.sourceType);if(n.length>0){let e=n.map(e=>`'${d(e,`sourceType`)}'`).join(`, `);t.push(`contentType IN (${e})`)}}if(e?.origin&&t.push(`origin = '${d(e.origin,`origin`)}'`),e?.category&&t.push(`category = '${d(e.category,`category`)}'`),e?.tags&&e.tags.length>0){let n=e.tags.map(e=>`tags LIKE '%${d(e,`tag`)}%'`);t.push(`(${n.join(` OR `)})`)}return t.length>0?t.join(` AND `):null}fromLanceRecord(e){return{id:e.id,content:e.content,sourcePath:e.sourcePath,contentType:e.contentType,headingPath:e.headingPath||void 0,chunkIndex:e.chunkIndex,totalChunks:e.totalChunks,startLine:e.startLine,endLine:e.endLine,fileHash:e.fileHash,indexedAt:e.indexedAt,origin:e.origin,tags:c(e.tags),category:e.category||void 0,version:e.version}}};export{f as LanceStore};