Self-learning is excluded from the community edition (ADR-0004); the project owner reworks it separately. Cost tracking stays. - agent: delete behaviour evolution (evolve-schema, iterative GEPA, compose-evolution, evolution gates, orchestrator and deploy), improvement and correction detection, AgentLearning, the eval-dataset builder, optimization-capture and the agent_insights tool. Secret detection moves from eval-dataset to secret-patterns. The LLM wiring keeps only the Haiku adapter and judge call used by skill audit and hygiene, so @ax-llm/ax is now declared by agent instead of server. - server: delete /api/evolution/*, the evolution and optimizer services, GEPA expansion of the first chat message, the prompt_optimization and monthly_assessment jobs, the idle-session reviewer and its session-reviewer persona, prompt injection of past corrections and thumbs-down recorded as correction signals. Existing job rows of the removed types are ignored, not deleted. - core: delete OptimizationLogStore; stop re-exporting the improvement signal and evolution run stores. - web: delete the Evolution screen and the two job types in the automation editor. - delete packages/optimizer. Also logs UF-027 (self-learning loops read data nothing collects) and UF-028 (the Settings prompt-shape choice is ignored by chat).
658 lines
30 KiB
TypeScript
658 lines
30 KiB
TypeScript
import type {
|
||
MindDB,
|
||
IdentityLayer,
|
||
AwarenessLayer,
|
||
FrameStore,
|
||
SessionStore,
|
||
HybridSearch,
|
||
KnowledgeGraph,
|
||
Embedder,
|
||
Importance,
|
||
FrameSource,
|
||
} from '@waggle/core';
|
||
import type { CognifyPipeline } from './cognify.js';
|
||
import type { FeedbackHandler } from './feedback-handler.js';
|
||
import type { RiskLevel } from '@waggle/shared';
|
||
import { createCoreLogger } from '@waggle/core';
|
||
import { detectContradiction } from './contradiction-detector.js';
|
||
import { scanForInjection } from './injection-scanner.js';
|
||
import { normalizeForDedup, cosineSimilarity, detectDramaticClaims, deriveConfidence, type ConfidenceLevel } from './text-analysis.js';
|
||
export type { ConfidenceLevel } from './text-analysis.js';
|
||
export { formatCombinedResult } from './result-formatter.js';
|
||
|
||
const log = createCoreLogger('mind-tools');
|
||
|
||
export interface ToolDefinition {
|
||
name: string;
|
||
description: string;
|
||
parameters: Record<string, unknown>;
|
||
execute: (args: Record<string, unknown>) => Promise<string>;
|
||
/**
|
||
* Provider-authored runtime risk. This is trusted metadata, never model input;
|
||
* confirmation policy may use it only to elevate name-based risk.
|
||
*/
|
||
riskLevel?: RiskLevel;
|
||
/** PM-6: Whether this tool can operate without LLM connectivity (default: false) */
|
||
offlineCapable?: boolean;
|
||
/**
|
||
* Dynamic availability check — called before each agent turn to decide
|
||
* whether this tool should be included in the tool list.
|
||
* Return false to hide the tool (e.g., browser not connected, no git repo).
|
||
* If not provided, the tool is always available.
|
||
*/
|
||
checkAvailability?: () => boolean;
|
||
}
|
||
|
||
/** Workspace-specific layers for dual-mind routing */
|
||
interface WorkspaceLayers {
|
||
db: MindDB;
|
||
frames: FrameStore;
|
||
sessions: SessionStore;
|
||
search: HybridSearch;
|
||
knowledge: KnowledgeGraph;
|
||
cognify: CognifyPipeline;
|
||
}
|
||
|
||
/** Tracks tool utilization across the current session */
|
||
export interface ToolUtilizationTracker {
|
||
/** Record that a tool was used */
|
||
recordUsage(toolName: string): void;
|
||
/** Get the set of unique tools used this session */
|
||
getUsedTools(): Set<string>;
|
||
/** Get the total number of available tools */
|
||
totalAvailable: number;
|
||
/** Get utilization ratio (uniqueUsed / totalAvailable) */
|
||
getUtilization(): number;
|
||
}
|
||
|
||
/** Create a ToolUtilizationTracker instance */
|
||
export function createToolUtilizationTracker(totalAvailable: number): ToolUtilizationTracker {
|
||
const usedTools = new Set<string>();
|
||
return {
|
||
recordUsage(toolName: string) {
|
||
usedTools.add(toolName);
|
||
},
|
||
getUsedTools() {
|
||
return new Set(usedTools);
|
||
},
|
||
totalAvailable,
|
||
getUtilization() {
|
||
if (totalAvailable === 0) return 0;
|
||
return usedTools.size / totalAvailable;
|
||
},
|
||
};
|
||
}
|
||
|
||
export interface MindToolDeps {
|
||
db: MindDB;
|
||
identity: IdentityLayer;
|
||
awareness: AwarenessLayer;
|
||
frames: FrameStore;
|
||
sessions: SessionStore;
|
||
search: HybridSearch;
|
||
knowledge: KnowledgeGraph;
|
||
cognify?: CognifyPipeline;
|
||
feedback?: FeedbackHandler;
|
||
/** F16: Optional embedder for semantic dedup (cosine similarity) */
|
||
embedder?: Embedder;
|
||
/** Dynamic accessor for workspace layers — checked at call time */
|
||
getWorkspaceLayers?: () => WorkspaceLayers | null;
|
||
/** W5.7: Accessor for ALL workspace search instances (for cross-workspace search) */
|
||
getAllWorkspaceSearches?: () => Array<{ workspaceId: string; workspaceName: string; search: HybridSearch }>;
|
||
/** Optional tool utilization tracker for session-level stats */
|
||
toolUtilizationTracker?: ToolUtilizationTracker;
|
||
/**
|
||
* Review C2: externally-owned save counter so the rate limit survives
|
||
* tool-array reconstruction (persona switch, workspace change, MCP reconnect).
|
||
* If omitted, a local counter is used (legacy behavior).
|
||
*/
|
||
saveCounter?: { count: number };
|
||
}
|
||
|
||
export function createMindTools(deps: MindToolDeps): ToolDefinition[] {
|
||
// W2.10: Per-session memory save rate limiter (prevents memory flooding attacks)
|
||
// Review C2: use externally-owned counter when provided so limit survives
|
||
// tool-array reconstruction (persona switch, workspace change, MCP reconnect).
|
||
const MAX_SAVES_PER_SESSION = 50;
|
||
const saveCounter = deps.saveCounter ?? { count: 0 };
|
||
|
||
return [
|
||
{
|
||
name: 'get_identity',
|
||
description: 'Get the agent identity (who am I, what is my role)',
|
||
offlineCapable: true,
|
||
parameters: {},
|
||
execute: async () => {
|
||
if (!deps.identity.exists()) return 'No identity configured yet.';
|
||
return deps.identity.toContext();
|
||
},
|
||
},
|
||
{
|
||
name: 'get_awareness',
|
||
description: 'Get current awareness state (active tasks, recent actions, pending items, flags, tool utilization)',
|
||
offlineCapable: true,
|
||
parameters: {},
|
||
execute: async () => {
|
||
let context = deps.awareness.toContext();
|
||
|
||
// Append utilization stats if tracker is available
|
||
if (deps.toolUtilizationTracker) {
|
||
const tracker = deps.toolUtilizationTracker;
|
||
const used = tracker.getUsedTools();
|
||
const utilization = tracker.getUtilization();
|
||
context += `\n\n## Tool Utilization\n`;
|
||
context += `- Unique tools used this session: ${used.size}\n`;
|
||
context += `- Total available tools: ${tracker.totalAvailable}\n`;
|
||
context += `- Utilization: ${(utilization * 100).toFixed(1)}%\n`;
|
||
if (utilization < 0.1) {
|
||
context += `- Note: Low utilization — consider using more of your available tools (skills, workflows, sub-agents) for richer results.\n`;
|
||
}
|
||
}
|
||
|
||
return context;
|
||
},
|
||
},
|
||
{
|
||
name: 'search_memory',
|
||
description: 'Search through memory for relevant past experiences and knowledge. Searches both personal and workspace memory when a workspace is active.',
|
||
offlineCapable: true,
|
||
parameters: {
|
||
type: 'object',
|
||
properties: {
|
||
query: { type: 'string', description: 'What to search for' },
|
||
profile: {
|
||
type: 'string',
|
||
enum: ['balanced', 'recent', 'important', 'connected'],
|
||
description: 'Scoring profile to use',
|
||
},
|
||
scope: {
|
||
type: 'string',
|
||
enum: ['all', 'personal', 'workspace'],
|
||
description: 'Which mind to search (default: all)',
|
||
},
|
||
limit: { type: 'number', description: 'Max results to return' },
|
||
importance: {
|
||
type: 'string',
|
||
enum: ['critical', 'important', 'normal', 'temporary', 'all'],
|
||
description: 'Filter by importance level (default: all)',
|
||
},
|
||
since: {
|
||
type: 'string',
|
||
description: 'ISO date — only return frames created after this date (e.g., "2026-03-14")',
|
||
},
|
||
until: {
|
||
type: 'string',
|
||
description: 'ISO date — only return frames created before this date (e.g., "2026-03-21")',
|
||
},
|
||
},
|
||
required: ['query'],
|
||
},
|
||
execute: async (args) => {
|
||
const limit = (args.limit as number) ?? 10;
|
||
const scope = (args.scope as string) ?? 'all';
|
||
const query = args.query as string;
|
||
const profile = (args.profile as 'balanced') ?? 'balanced';
|
||
const importanceFilter = (args.importance as string) ?? 'all';
|
||
// F20: Temporal filtering params
|
||
const since = args.since as string | undefined;
|
||
const until = args.until as string | undefined;
|
||
const wsLayers = deps.getWorkspaceLayers?.();
|
||
|
||
const sections: string[] = [];
|
||
|
||
// F14: Helper to filter results by importance level
|
||
const filterByImportance = <T extends { frame: { importance: string } }>(results: T[]): T[] => {
|
||
if (importanceFilter === 'all') return results;
|
||
return results.filter(r => r.frame.importance === importanceFilter);
|
||
};
|
||
|
||
// W2.6: Search workspace mind FIRST (results shown before personal, boosted priority)
|
||
if (wsLayers && (scope === 'all' || scope === 'workspace')) {
|
||
const wsResults = filterByImportance(await wsLayers.search.search(query, { limit, profile, since, until }));
|
||
if (wsResults.length > 0) {
|
||
sections.push('## Workspace Memory');
|
||
sections.push(...wsResults.map((r, i) =>
|
||
`[${i + 1}] (score: ${r.finalScore.toFixed(3)}, date: ${r.frame.created_at?.slice(0, 10) ?? 'unknown'}, type: ${r.frame.frame_type}, importance: ${r.frame.importance}, source: ${r.frame.source ?? 'user_stated'})\n${r.frame.content}`
|
||
));
|
||
}
|
||
}
|
||
|
||
// Search personal mind (shown after workspace results, with reduced limit when workspace is active)
|
||
if (scope === 'all' || scope === 'personal') {
|
||
// W2.6: When inside a workspace, cap personal results to reduce noise
|
||
const personalLimit = (wsLayers && scope === 'all') ? Math.min(limit, 3) : limit;
|
||
const personalResults = filterByImportance(await deps.search.search(query, { limit: personalLimit, profile, since, until }));
|
||
if (personalResults.length > 0) {
|
||
if (wsLayers) {
|
||
sections.push('## Personal Memory\n_(Cross-workspace — not specific to this workspace)_');
|
||
}
|
||
sections.push(...personalResults.map((r, i) =>
|
||
`[${i + 1}] (score: ${r.finalScore.toFixed(3)}, date: ${r.frame.created_at?.slice(0, 10) ?? 'unknown'}, type: ${r.frame.frame_type}, importance: ${r.frame.importance}, source: ${r.frame.source ?? 'user_stated'})\n${r.frame.content}`
|
||
));
|
||
}
|
||
}
|
||
|
||
// W2.5: Detect potential contradictions in results
|
||
if (sections.length > 1) {
|
||
const allContent = sections.filter(s => !s.startsWith('##')).join('\n');
|
||
const contradictionSignals = ['but ', 'however', 'instead', 'no longer', 'changed to', 'reversed', 'cancelled', 'not '];
|
||
const hasConflictSignal = contradictionSignals.some(s => allContent.toLowerCase().includes(s));
|
||
if (hasConflictSignal) {
|
||
sections.push('\n⚠️ Note: Multiple memories found. Check for potential contradictions — verify which is the most current/authoritative version.');
|
||
}
|
||
}
|
||
|
||
if (sections.length > 0) return sections.join('\n\n');
|
||
|
||
// Fallback: LIKE scan on both minds
|
||
const allFallback: string[] = [];
|
||
const keywords = query.toLowerCase().split(/\s+/).filter(w => w.length > 2);
|
||
|
||
for (const [label, db] of [['personal', deps.db], ...(wsLayers ? [['workspace', wsLayers.db] as const] : [])] as const) {
|
||
const raw = db.getDatabase();
|
||
type FallbackFrameRow = { id: number; content: string; frame_type: string; importance: string };
|
||
let fallbackFrames: FallbackFrameRow[];
|
||
if (keywords.length > 0) {
|
||
const likeClauses = keywords.map(() => "LOWER(content) LIKE '%' || ? || '%' ESCAPE '\\'").join(' OR ');
|
||
const likeParams = keywords.map(k => k.replace(/\\/g, '\\\\').replace(/%/g, '\\%').replace(/_/g, '\\_'));
|
||
fallbackFrames = raw.prepare(
|
||
`SELECT id, content, frame_type, importance FROM memory_frames WHERE ${likeClauses} ORDER BY id DESC LIMIT ?`
|
||
).all(...likeParams, limit) as FallbackFrameRow[];
|
||
} else {
|
||
fallbackFrames = raw.prepare(
|
||
'SELECT id, content, frame_type, importance FROM memory_frames ORDER BY id DESC LIMIT ?'
|
||
).all(limit) as FallbackFrameRow[];
|
||
}
|
||
if (fallbackFrames.length > 0) {
|
||
if (wsLayers) allFallback.push(`## ${label === 'workspace' ? 'Workspace' : 'Personal'} Memory`);
|
||
allFallback.push(...fallbackFrames.map((f, i) =>
|
||
`[${i + 1}] (type: ${f.frame_type}, importance: ${f.importance})\n${f.content}`
|
||
));
|
||
}
|
||
}
|
||
|
||
if (allFallback.length === 0) return 'No relevant memories found.';
|
||
return allFallback.join('\n\n');
|
||
},
|
||
},
|
||
// W5.7: Cross-workspace search — searches ALL workspace minds + personal
|
||
{
|
||
name: 'search_all_workspaces',
|
||
description: 'Search across ALL workspaces and personal memory simultaneously. Use for cross-project queries like "what should I work on?" or "find everything about topic X across all projects".',
|
||
offlineCapable: true,
|
||
parameters: {
|
||
type: 'object',
|
||
properties: {
|
||
query: { type: 'string', description: 'What to search for across all workspaces' },
|
||
limit: { type: 'number', description: 'Max results per workspace (default: 5)' },
|
||
},
|
||
required: ['query'],
|
||
},
|
||
execute: async (args) => {
|
||
const query = args.query as string;
|
||
const limitPerWs = (args.limit as number) ?? 5;
|
||
const sections: string[] = [];
|
||
|
||
// Search personal mind
|
||
const personalResults = await deps.search.search(query, { limit: limitPerWs, profile: 'balanced' });
|
||
if (personalResults.length > 0) {
|
||
sections.push('## Personal Memory');
|
||
sections.push(...personalResults.map((r, i) =>
|
||
`[${i + 1}] (score: ${r.finalScore.toFixed(3)}, importance: ${r.frame.importance})\n${r.frame.content}`
|
||
));
|
||
}
|
||
|
||
// Search all workspace minds
|
||
const allWs = deps.getAllWorkspaceSearches?.() ?? [];
|
||
for (const { workspaceId, workspaceName, search: wsSearch } of allWs) {
|
||
try {
|
||
const wsResults = await wsSearch.search(query, { limit: limitPerWs, profile: 'balanced' });
|
||
if (wsResults.length > 0) {
|
||
sections.push(`## Workspace: ${workspaceName} (${workspaceId})`);
|
||
sections.push(...wsResults.map((r, i) =>
|
||
`[${i + 1}] (score: ${r.finalScore.toFixed(3)}, importance: ${r.frame.importance})\n${r.frame.content}`
|
||
));
|
||
}
|
||
} catch (err) { log.warn(`search_all_workspaces: skipped workspace ${workspaceId}`, err); }
|
||
}
|
||
|
||
if (sections.length === 0) return 'No relevant memories found across any workspace.';
|
||
return sections.join('\n\n');
|
||
},
|
||
},
|
||
{
|
||
name: 'save_memory',
|
||
description: 'Save a new memory. Defaults to workspace mind when active.\n\n'
|
||
+ 'CRITICAL ROUTING RULES:\n'
|
||
+ '- target="workspace" (DEFAULT): ALL project data, client data, decisions, tasks, meeting notes, research findings, confidential information, names, budgets, timelines. '
|
||
+ 'Workspace memories are ISOLATED — they cannot be seen from other workspaces. Use this for ANYTHING workspace-specific.\n'
|
||
+ '- target="personal": ONLY for user preferences (favorite color, communication style), personal facts (name, role, timezone), '
|
||
+ 'and cross-workspace knowledge (general skills, tool preferences). Personal memories are shared across ALL workspaces.\n\n'
|
||
+ 'When in doubt, use workspace. Never save client names, budgets, confidential data, or project-specific information to personal mind.',
|
||
offlineCapable: true,
|
||
parameters: {
|
||
type: 'object',
|
||
properties: {
|
||
content: { type: 'string', description: 'The memory content to save' },
|
||
importance: {
|
||
type: 'string',
|
||
enum: ['critical', 'important', 'normal', 'temporary'],
|
||
description: 'How important is this memory',
|
||
},
|
||
target: {
|
||
type: 'string',
|
||
enum: ['workspace', 'personal'],
|
||
description: 'Which mind to save to. Workspace (default): ALL project/client/confidential data — isolated per workspace. Personal: ONLY user preferences and cross-workspace knowledge — shared across all workspaces.',
|
||
},
|
||
source: {
|
||
type: 'string',
|
||
enum: ['user_stated', 'tool_verified', 'agent_inferred'],
|
||
description: 'Provenance: user_stated = the user said this, tool_verified = confirmed via tool output, agent_inferred = your own analysis/synthesis.',
|
||
},
|
||
confidence: {
|
||
type: 'string',
|
||
enum: ['high', 'medium', 'low', 'unverified'],
|
||
description: 'Confidence level. Auto-derived from source if omitted: tool_verified→high, user_stated→medium, agent_inferred→low.',
|
||
},
|
||
},
|
||
required: ['content'],
|
||
},
|
||
execute: async (args) => {
|
||
const content = args.content as string;
|
||
|
||
// Review C1: scan LLM-controlled content for prompt-injection payloads
|
||
// BEFORE any DB write. The pre:memory-write hook is a cancellation gate,
|
||
// not a scanner — if no hook is registered, content goes straight to SQLite.
|
||
const scan = scanForInjection(content, 'user_input');
|
||
if (!scan.safe) {
|
||
return `Memory save blocked: content contains suspected prompt-injection payload (${scan.flags.join(', ')}). The content was NOT saved.`;
|
||
}
|
||
|
||
// W2.10: Rate limit memory saves to prevent flooding (50 per session)
|
||
// Review C2: counter is externally-owned so it survives tool-array rebuilds.
|
||
saveCounter.count++;
|
||
if (saveCounter.count > MAX_SAVES_PER_SESSION) {
|
||
return `Memory save rate limit reached (${MAX_SAVES_PER_SESSION} saves this session). This prevents memory flooding. Start a new session to save more memories.`;
|
||
}
|
||
|
||
// Untrusted args: validate against the canonical unions, default on mismatch.
|
||
const VALID_IMPORTANCE: readonly Importance[] = ['critical', 'important', 'normal', 'temporary', 'deprecated'];
|
||
const VALID_SOURCE: readonly FrameSource[] = ['user_stated', 'tool_verified', 'agent_inferred', 'import', 'system'];
|
||
let importance: Importance =
|
||
typeof args.importance === 'string' && (VALID_IMPORTANCE as readonly string[]).includes(args.importance)
|
||
? (args.importance as Importance)
|
||
: 'normal';
|
||
const target = typeof args.target === 'string' ? args.target : 'workspace';
|
||
const source: FrameSource =
|
||
typeof args.source === 'string' && (VALID_SOURCE as readonly string[]).includes(args.source)
|
||
? (args.source as FrameSource)
|
||
: 'user_stated';
|
||
const wsLayers = deps.getWorkspaceLayers?.();
|
||
|
||
// F6: Derive confidence from source if not explicitly provided
|
||
const confidence: ConfidenceLevel = (args.confidence as ConfidenceLevel) ?? deriveConfidence(source);
|
||
|
||
// F7: Strict workspace routing — respect target parameter, with safety guardrail
|
||
// When target='workspace' AND workspace is active → save to workspace only
|
||
// When target='personal' → save to personal, UNLESS content appears workspace-specific
|
||
let effectiveTarget = target;
|
||
|
||
// B1-guardrail: Detect workspace-specific content being routed to personal mind
|
||
// If workspace is active and content contains confidential/project signals, redirect to workspace
|
||
if (effectiveTarget === 'personal' && wsLayers != null) {
|
||
const workspaceSpecificPatterns = [
|
||
/\bconfidential\b/i, /\bclient\b/i, /\bbudget\b/i, /\bengagement\b/i,
|
||
/\bdeliverable\b/i, /\bmilestone\b/i, /\btimeline\b/i, /\bdeadline\b/i,
|
||
/\bceo\b/i, /\bcto\b/i, /\bcfo\b/i, /\bcoo\b/i,
|
||
/\$\d{2,}[,.]?\d*[kmb]?\b/i, // dollar amounts ($500K, $1.2M, etc.)
|
||
/\bproject\b.*\b(?:plan|status|update)\b/i,
|
||
/\bcontract\b/i, /\bproposal\b/i, /\bnda\b/i, /\bsow\b/i,
|
||
];
|
||
const hasWorkspaceSignal = workspaceSpecificPatterns.some(p => p.test(content));
|
||
if (hasWorkspaceSignal) {
|
||
effectiveTarget = 'workspace';
|
||
}
|
||
}
|
||
|
||
const useWorkspace = effectiveTarget === 'workspace' && wsLayers != null;
|
||
const targetFrames = useWorkspace ? wsLayers!.frames : deps.frames;
|
||
const targetSessions = useWorkspace ? wsLayers!.sessions : deps.sessions;
|
||
const targetCognify = useWorkspace ? wsLayers!.cognify : deps.cognify;
|
||
const targetDb = useWorkspace ? wsLayers!.db : deps.db;
|
||
const mindLabel = useWorkspace ? 'workspace' : 'personal';
|
||
|
||
// F22: Detect dramatic claims and flag them
|
||
const dramaticClaims = detectDramaticClaims(content);
|
||
let dramaticFlag = '';
|
||
if (dramaticClaims.length) {
|
||
// Downgrade importance to 'temporary' for dramatic claims unless explicitly set to critical
|
||
if (importance !== 'critical') {
|
||
importance = 'temporary';
|
||
}
|
||
dramaticFlag = ` [flag: dramatic_claim (${dramaticClaims.join(', ')})]`;
|
||
}
|
||
|
||
// ── Dedup checks (F7 cross-mind + F16 enhanced) ──────────────
|
||
|
||
// Helper: check a single database for exact or normalized duplicates
|
||
const checkDedupInDb = async (db: MindDB, label: string): Promise<string | null> => {
|
||
try {
|
||
const raw = db.getDatabase();
|
||
|
||
// Exact-match dedup (original W2.4 check)
|
||
const exactMatch = raw.prepare(
|
||
"SELECT id, content FROM memory_frames WHERE content = ? LIMIT 1"
|
||
).get(content) as { id: number; content: string } | undefined;
|
||
if (exactMatch) {
|
||
return `Memory already exists in ${label} mind (frame ${exactMatch.id}) — skipped duplicate save.`;
|
||
}
|
||
|
||
// F16: Normalized string comparison
|
||
const normalizedContent = normalizeForDedup(content);
|
||
if (normalizedContent.length > 0) {
|
||
// Fetch recent frames for normalized comparison (cap at 100 to avoid perf issues)
|
||
const recentFrames = raw.prepare(
|
||
"SELECT id, content FROM memory_frames ORDER BY id DESC LIMIT 100"
|
||
).all() as Array<{ id: number; content: string }>;
|
||
|
||
for (const frame of recentFrames) {
|
||
if (normalizeForDedup(frame.content) === normalizedContent) {
|
||
return `Memory already exists in ${label} mind (frame ${frame.id}, normalized match) — skipped duplicate save.`;
|
||
}
|
||
}
|
||
}
|
||
|
||
// F16: Embedding-based similarity dedup (if embedder available)
|
||
if (deps.embedder && normalizedContent.length > 0) {
|
||
try {
|
||
const candidateFrames = raw.prepare(
|
||
"SELECT id, content FROM memory_frames ORDER BY id DESC LIMIT 50"
|
||
).all() as Array<{ id: number; content: string }>;
|
||
|
||
if (candidateFrames.length > 0) {
|
||
const newEmbedding = await deps.embedder.embed(content);
|
||
const candidateTexts = candidateFrames.map(f => f.content);
|
||
const candidateEmbeddings = await deps.embedder.embedBatch(candidateTexts);
|
||
|
||
for (let i = 0; i < candidateEmbeddings.length; i++) {
|
||
const similarity = cosineSimilarity(newEmbedding, candidateEmbeddings[i]);
|
||
if (similarity > 0.95) {
|
||
return `Memory already exists in ${label} mind (frame ${candidateFrames[i].id}, semantic similarity: ${similarity.toFixed(3)}) — skipped duplicate save.`;
|
||
}
|
||
}
|
||
}
|
||
} catch { /* non-blocking — continue if embedding dedup fails */ }
|
||
}
|
||
} catch { /* non-blocking — proceed with save if dedup check fails */ }
|
||
return null;
|
||
};
|
||
|
||
// F7: Cross-mind dedup — check BOTH workspace and personal mind before saving
|
||
const targetDedupResult = await checkDedupInDb(targetDb, mindLabel);
|
||
if (targetDedupResult) return targetDedupResult;
|
||
|
||
// Check the OTHER mind for cross-mind duplicates
|
||
if (useWorkspace) {
|
||
// Saving to workspace — also check personal mind
|
||
const crossDedupResult = await checkDedupInDb(deps.db, 'personal');
|
||
if (crossDedupResult) {
|
||
return crossDedupResult.replace('skipped duplicate save', 'skipped cross-mind duplicate save');
|
||
}
|
||
} else if (wsLayers) {
|
||
// Saving to personal — also check workspace mind
|
||
const crossDedupResult = await checkDedupInDb(wsLayers.db, 'workspace');
|
||
if (crossDedupResult) {
|
||
return crossDedupResult.replace('skipped duplicate save', 'skipped cross-mind duplicate save');
|
||
}
|
||
}
|
||
|
||
// Skills 2.0 gap K: write-path conflict detection.
|
||
// Check recent frames for contradictions BEFORE saving. When found,
|
||
// still save the new frame (conflicts are data, not errors) but
|
||
// emit a correction improvement signal and flag the response so
|
||
// the agent surfaces the conflict to the user.
|
||
let conflictFlag = '';
|
||
try {
|
||
const raw = targetDb.getDatabase();
|
||
const recentRows = raw.prepare(
|
||
"SELECT content FROM memory_frames ORDER BY id DESC LIMIT 50",
|
||
).all() as Array<{ content: string }>;
|
||
const contradiction = detectContradiction(content, recentRows);
|
||
if (contradiction.isContradiction && contradiction.conflictsWith) {
|
||
conflictFlag = ` [flag: contradicts_existing (${contradiction.conflictsWith.slice(0, 80).replace(/\s+/g, ' ').trim()}…)]`;
|
||
}
|
||
} catch { /* non-blocking — contradiction check failure proceeds with save */ }
|
||
|
||
// Use cognify pipeline if available (extracts entities + indexes for search)
|
||
if (targetCognify) {
|
||
// PR3.5 (review H-1): pass the agent's chosen source through cognify —
|
||
// dropping it stored every cognify-path memory as 'user_stated' while
|
||
// the response string still claimed the real source (a provenance lie).
|
||
const result = await targetCognify.cognify(content, importance, undefined, undefined, source);
|
||
return `Memory saved to ${mindLabel} mind (importance: ${importance}, source: ${source}, confidence: ${confidence}${dramaticFlag}${conflictFlag}, entities: ${result.entitiesExtracted}, relations: ${result.relationsCreated}).`;
|
||
}
|
||
|
||
// Fallback: raw frame creation (no entity extraction or vector indexing)
|
||
const active = targetSessions.getActive();
|
||
let gopId: string;
|
||
if (active.length === 0) {
|
||
const session = targetSessions.create();
|
||
gopId = session.gop_id;
|
||
} else {
|
||
gopId = active[0].gop_id;
|
||
}
|
||
const latestI = targetFrames.getLatestIFrame(gopId);
|
||
if (latestI) {
|
||
targetFrames.createPFrame(gopId, content, latestI.id, importance, source);
|
||
} else {
|
||
targetFrames.createIFrame(gopId, content, importance, source);
|
||
}
|
||
return `Memory saved to ${mindLabel} mind (importance: ${importance}, source: ${source}, confidence: ${confidence}${dramaticFlag}${conflictFlag}).`;
|
||
},
|
||
},
|
||
{
|
||
name: 'query_knowledge',
|
||
description: 'Query the knowledge graph for entities and their relationships. Use query="*" to list all known entities.',
|
||
offlineCapable: true,
|
||
parameters: {
|
||
type: 'object',
|
||
properties: {
|
||
query: { type: 'string', description: 'Entity name to search for, or "*" to list all entities' },
|
||
type: { type: 'string', description: 'Filter by entity type (e.g., person, project, technology, concept, organization)' },
|
||
},
|
||
required: ['query'],
|
||
},
|
||
execute: async (args) => {
|
||
const query = args.query as string;
|
||
const entityType = args.type as string | undefined;
|
||
|
||
let entities;
|
||
// F5: Wildcard support — list all entities
|
||
if (query === '*' || query.toLowerCase() === 'all') {
|
||
entities = deps.knowledge.getEntitiesByType(entityType ?? '');
|
||
} else if (entityType) {
|
||
entities = deps.knowledge.getEntitiesByType(entityType);
|
||
entities = entities.filter(e => e.name.toLowerCase().includes(query.toLowerCase()));
|
||
} else {
|
||
entities = deps.knowledge.searchEntities(query);
|
||
}
|
||
|
||
if (entities.length === 0) return `No entities found matching "${query}".`;
|
||
|
||
const maxEntities = (query === '*' || query.toLowerCase() === 'all') ? 50 : 10;
|
||
const results: string[] = [];
|
||
for (const entity of entities.slice(0, maxEntities)) {
|
||
const rels = deps.knowledge.getRelationsFrom(entity.id);
|
||
const relStr = rels.length > 0
|
||
? rels.map(r => {
|
||
const target = deps.knowledge.getEntity(r.target_id);
|
||
return ` → ${r.relation_type} → ${target?.name ?? 'unknown'} (confidence: ${r.confidence})`;
|
||
}).join('\n')
|
||
: ' (no outgoing relations)';
|
||
results.push(`${entity.entity_type}: ${entity.name}\n${relStr}`);
|
||
}
|
||
return results.join('\n\n');
|
||
},
|
||
},
|
||
{
|
||
name: 'add_task',
|
||
description: 'Add a task to the awareness layer for tracking',
|
||
offlineCapable: true,
|
||
parameters: {
|
||
type: 'object',
|
||
properties: {
|
||
content: { type: 'string', description: 'Task description' },
|
||
priority: { type: 'number', description: 'Priority (0-10, higher = more important)' },
|
||
},
|
||
required: ['content'],
|
||
},
|
||
execute: async (args) => {
|
||
const taskContent = typeof args.content === 'string' ? args.content.slice(0, 2000) : '';
|
||
const item = deps.awareness.add('task', taskContent, (args.priority as number) ?? 5);
|
||
return `Task added: "${item.content}" (priority: ${item.priority})`;
|
||
},
|
||
},
|
||
{
|
||
name: 'correct_knowledge',
|
||
description: 'Correct or invalidate a knowledge entity. Use entity_id from query_knowledge results.',
|
||
parameters: {
|
||
type: 'object',
|
||
properties: {
|
||
entity_id: { type: 'number', description: 'The entity ID to update' },
|
||
action: {
|
||
type: 'string',
|
||
enum: ['correct', 'invalidate'],
|
||
description: 'Whether to correct or invalidate the entity',
|
||
},
|
||
updates: {
|
||
type: 'object',
|
||
description: 'For correct: property updates to merge. For invalidate: { reason: "why it is wrong" }',
|
||
},
|
||
},
|
||
required: ['entity_id', 'action'],
|
||
},
|
||
execute: async (args) => {
|
||
if (!deps.feedback) {
|
||
return 'Feedback handler not available. Cannot correct knowledge.';
|
||
}
|
||
|
||
const entityId = args.entity_id as number;
|
||
const action = args.action as string;
|
||
const updates = (args.updates as Record<string, unknown>) ?? {};
|
||
|
||
if (action === 'correct') {
|
||
deps.feedback.correctEntity(entityId, updates);
|
||
return `Entity ${entityId} corrected with updates: ${JSON.stringify(updates)}`;
|
||
} else if (action === 'invalidate') {
|
||
const reason = (updates.reason as string) ?? 'No reason provided';
|
||
deps.feedback.invalidateEntity(entityId, reason);
|
||
return `Entity ${entityId} invalidated. Reason: ${reason}`;
|
||
}
|
||
|
||
return `Unknown action "${action}". Use "correct" or "invalidate".`;
|
||
},
|
||
},
|
||
];
|
||
}
|
||
|