orionarchitekton
64 verified skills64 total stars
skills/oia/gtm-engine
# gtm-engine **Lane:** oia **Version:** 0.1.0 **Owner:** system **Schedule:** 59 23 29 2 * ## Purpose Complete Phase 1 GTM pack generation from business idea to deliverables. Runs prompts 1-10 with model routing (Edge for cloud, Atlas for local GPU). Produces: Executive Summary, Business Plan, 90-Day Launch Plan, 5 Playbooks. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ##
tools1
skills/oia/gtm-feedback-loop
# gtm-feedback-loop **Lane:** oia **Version:** 0.1.0 **Owner:** system **Schedule:** 0 2 * * 0 ## Purpose Weekly analysis of live GTM performance data (KPIs, marketing, finance). Runs on Atlas with deepseek-r1:7b for cost-effective local reasoning. Produces: Variance Report, Recommendations, Automation Hooks. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Ever
tools1
skills/oia/gtm-omega-weekly
# gtm-omega-weekly **Lane:** oia **Version:** 0.1.0 **Owner:** system **Schedule:** 0 6 * * 0 ## Purpose Weekly GTM health check for Cosmocrat. Runs Layers 0-3.5 (Context, Market, Business Model, GTM Execution, Communication) plus Block P (Drift Detection). Provides comprehensive GTM readiness assessment. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every wr
tools1
skills/oia/gtm-self-audit
# gtm-self-audit **Lane:** oia **Version:** 0.1.0 **Owner:** system **Schedule:** 0 3 1 */3 * ## Purpose Quarterly audit of all GTM Engine components for data drift and staleness. Runs on Atlas with deepseek-r1:7b. Produces: Integrity Table, Drift Matrix, Regeneration Plan. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source
tools1
skills/oia/omega
# omega **Lane:** oia **Version:** 0.1.0 **Owner:** system **Schedule:** manual only ## Purpose Shared OMEGA GTM synthesis runner used by the cadence jobs. It executes the layered OMEGA pack with checkpointing, artifact generation, and gate evaluation, while allowing cadence wrappers to constrain the active layer and block scope. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None to MemoryClient.
tools1
skills/ops/agent-seos
# agent-seos **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 54 23 * * * ## Purpose Generates SEO recommendations from URLs and content topics mentioned in daily conversations. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | From Langfuse o
tools1
skills/ops/api-runner
# api-runner **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 28 0 * * * ## Purpose Executes a list of API tasks, logs execution to Langfuse, and persists events/artifacts to ClickHouse. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | From L
tools1
skills/ops/chronicle-ingest
# chronicle-ingest **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 40 0 * * * ## Purpose Fetches recent Chronicle messages and ingests them to ClickHouse + pgvector. This ensures Chronicle posts are searchable in the unified memory system. Runs after the daily agents complete. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes
tools1
skills/ops/context-query-rewrite
# Context Query Rewrite Skill Rewrites user queries for better semantic retrieval by: - Extracting intent from natural language - Adding lane/domain filters automatically - Inferring temporal windows (recent vs. all-time) ## Phase **pre_retrieval** - Runs before semantic search retrieval ## Priority 10 (runs early in pre-retrieval phase) ## Activation - **Lanes**: ops (via `lane` field) - **Modes**: COO, CFO (via `activation.modes`) - **Agent Types**: cursor, codex (via `activation.agent_
development1
skills/ops/coo-action-engine-plan
# coo-action-engine-plan **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 18 9 * * * ## Purpose Reads COO Briefing and produces a task queue for the day. Queue is written to /cosmocrat/action_engine/queue/{DATE}.json ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generate
tools1
skills/ops/coo-alerts
# coo-alerts **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** */10 * * * * ## Purpose Monitors system health every 10 minutes. Alerts on: ingest failures, agent issues, ClickHouse/Langfuse connectivity, disk space. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated
tools1
skills/ops/daily-scout
# daily-scout **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 18 0 * * * ## Purpose Comprehensive daily health check: verifies all conversation sources, agent outputs, memory tables, Langfuse connectivity, and executive reports. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_i
tools1
skills/ops/decision-extractor-daily
# decision-extractor-daily **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 12 4 * * * ## Purpose Extracts decisions from conversations and populates decision_log. Runs after memory consolidation. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id
tools1
skills/ops/governance-agent-detection
# governance-agent-detection **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 30 0 * * * ## Purpose Auto-detects new agents from agent_events table and updates agents.yaml. Scans for new agent_id values, compares against agents.yaml, and auto-updates with empty allowlist (fail-closed safe). Runs nightly after agent events are logged. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes No
tools1
skills/ops/governance-registry-sync
# governance-registry-sync **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 5 6,18 * * * ## Purpose Auto-discovers new ClickHouse tables and updates registry.yaml. Runs 5 minutes after context-refresh to discover new tables, post to Chronicle, and auto-commit updates to cosmocrat1 branch. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every wri
tools1
skills/ops/knowledge-refresh
# knowledge-refresh **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 22 0 * * * ## Purpose Normalizes Processed_Knowledge exports and loads the results into cosmocrat.orion_knowledge so Truth Engine always has up-to-date long-term context. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------
tools1
skills/ops/postgres-ttl-cleanup
# postgres-ttl-cleanup **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 0 3 * * * ## Purpose Daily cleanup of expired rows from Postgres tables lacking native TTL. Deletes old audit logs (90d), idempotency keys, workspace invites (30d), subscriber events (60d). Logs deletions to ClickHouse audit trail. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provena
tools1
skills/ops/program-gate-check-integration-tests
# program-gate-check-integration-tests **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** manual ## Purpose Validates that integration tests pass (Phase 1 gate). ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | From Langfuse or run_id | | `sour
tools1
skills/ops/scheduler-rush-validation
# scheduler-rush-validation **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 30 2 * * * ## Purpose Automatic validation of scheduler performance after nightly rush hour (23:00-01:00 PT). Checks for critical delays, timeout issues, runs Prometheus scorecard, and posts results to Chronicle. Self-validates and reports via Slack #orion-dev. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes
tools1
skills/ops/validate-seos
# validate-seos **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 59 23 * * * ## Purpose Validates SEOS agent output quality and performance. Final validation job - triggers progression check after all agents validated. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Genera
tools1
skills/ops/workspace-manifest-attestor
# workspace-manifest-attestor **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** manual ## Purpose Attests workspace integrity by comparing discovered git roots to `pandora.repos.json`. Emits `WORKSPACE_MANIFEST_ATTESTOR` receipts with `missing_from_manifest`, `unexpected_git_roots`, and `risk_score`. ## Memory Contract ### Access Mode **`service_only`** - No direct DB writes. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source |
testing1
skills/personal/life-events-daily
# life-events-daily **Lane:** personal **Version:** 0.1.0 **Owner:** system **Schedule:** 15 3 * * * ## Purpose Extracts significant life events from the past week of messages using LLM analysis. Populates cosmocrat.life_events table. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Genera
tools1
skills/tbm/tbm-auto-heal
# tbm-auto-heal **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 27 * * * * ## Purpose Hourly ClickHouse + TBM healthcheck on atlas-01 (native port 9000) to fail fast on ingest issues. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | From Lan
tools1
skills/tbm/tbm-backfill
# tbm-backfill **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 3 * * * * ## Purpose Hourly rate-limited ingestion of TikTok/YouTube content. Fetches 3-5 videos per hour to avoid rate limits on Atlas. Tracks progress in tbm_ingest_progress table. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|-
tools1
skills/tbm/tbm-reading-sync
# tbm-reading-sync **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 30 22 * * * ## Purpose Daily sync of tiktok_transcripts to tbm_readings. Uses rule-based extraction for cards, themes, and spread type. Runs after backfill has completed for the day. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-----
tools1
skills/tbm/tiktok-overnight-pipeline
# tiktok-overnight-pipeline **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 0 1 * * * ## Purpose Complete overnight pipeline for 3-channel TikTok content generation. Generates 9 videos (3 per channel) with TikTok Shop affiliate products. Timeline: 01:00 - 06:10 PT with GPU management (Llama on/off cycles). ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Pr
tools1
skills/memory/update-memory-snapshot
# update-memory-snapshot ## Purpose Consolidates daily reports and agent outputs into long-term memory structures: - **Episodic Memory** - Short-term working memory with 30-day TTL - **Decision Traces** - Audit trail of successful patterns - **Lane Profiles** - Consolidated playbooks per vertical ## What It Does 1. **Reads** agent outputs from `events_daily` and `artifacts_daily` for the date 2. **Extracts** facts, decisions, and patterns from CMA/PAR/SEOS outputs 3. **Transforms** into stru
testing1
skills/oia/gtm-omega-monthly
# gtm-omega-monthly **Lane:** oia **Version:** 0.1.0 **Owner:** system **Schedule:** 0 6 1 * * ## Purpose Monthly full capital-ready GTM analysis for Cosmocrat. Runs all 24 blocks across all layers including Capital Intelligence (L4) and IP/Licensing Governance (L4.5). Produces complete launch artifact suite. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Ever
tools1
skills/ops/agent-par
# agent-par **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 48 23 * * * ## Purpose Extracts financial insights and profit optimization opportunities from conversations and executive reports. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | F
tools1
skills/ops/alexa
# alexa **Lane:** ops **Version:** 0.1.0 **Owner:** system ## Purpose Provide a queued execution path for Alexa-originated operations that exceed voice-response deadlines. ## Entry Point ```bash python -m app.services.mcp.jobs.alexa.job_worker ``` ## Contracts - Queue names: - `alexa:job_queue` - `alexa:job_queue:dlq` - `alexa:job_queue:poison` - Invalid payloads route to poison queue. - Retryable handler failures route to DLQ. ## Environment - `REDIS_URL` - `ALEXA_WORKER_METR
tools1
skills/ops/canary-post-validation
# canary-post-validation **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 15 6,18 * * * ## Purpose Automated validation after context-refresh completes. Validates stage metrics, audit log, pass/fail criteria. Posts results to Chronicle for team visibility. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source |
tools1
skills/ops/canary-pre-flight
# canary-pre-flight **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 30 5,17 * * * ## Purpose Automated pre-flight validation before context-refresh runs. Checks ClickHouse connectivity, audit table, manifest config, disk space. Includes governance config validation (check #11). Runs 30 minutes before each context-refresh. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-o
tools1
skills/ops/context-pack
# Context Pack Skill Packs retrieved context intelligently by: - Prioritizing decisions/invariants over chat history - Allocating token budget: must-include → nice-to-include → drop - Detecting conflicts between retrieved items - Attaching inclusion reasons for each item (for trace/replay) ## Phase **context_pack** - Runs after retrieval-time skills, before model call ## Priority 20 (runs after retrieval-time skills) ## Dependencies - `context-query-rewrite` (optional, if needs rewritten
testing1
skills/ops/context-refresh-metrics
# context-refresh-metrics **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 18 * * * * ## Purpose Exports context refresh metrics to Prometheus textfile collector. Tracks: last success timestamp, duration, consecutive CH failures. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_i
tools1
skills/ops/context-refresh
# context-refresh **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 0 6,18 * * * ## Purpose Twice-daily synthesis of context files from telemetry sources: - Chronicle thread (ChatGPT task summaries) - ClickHouse (events_daily, artifacts_daily, agent_events) - Git (recent commits and changed files) Production-hardened with process locking, incremental watermarks, atomic writes, and config-driven repos. ## Memory Contract ### Access Mode **`service_only`** - All writes
tools1
skills/ops/conversation-exporter
# conversation-exporter **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 35 23 * * * ## Purpose Export all ChatGPT conversations for the run date (America/Los_Angeles), normalize them, and write today_conversations JSON for the Executive Daily Report runner. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source
tools1
skills/ops/daily-report-verifier
# daily-report-verifier **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 12 0 * * * ## Purpose Checks that today's conversation export and Executive Daily Report outputs exist and that deltaComputed is consistent. Alerts Slack #ops on failure. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|----
tools1
skills/ops/import-conversation-logs
# import-conversation-logs **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 23:28 PT daily ## Purpose Ingest conversations from multiple AI chat platforms into the unified `messages_v2` table. This is the foundation of the memory pipeline - all other skills depend on this data. ## Sources | Platform | Inbox Path | Format | |----------|-----------|--------| | ChatGPT | `/cosmocrat/inbox/raw/chatgpt/` | OpenAI export JSON | | Claude | `/cosmocrat/inbox/raw/claude/` | Ant
development1
skills/ops/memory-consolidator
# memory-consolidator **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 45 1 * * * ## Purpose Nightly memory consolidation: exports Langfuse traces, summarizes decision traces into lane profiles, and prunes expired episodic memory. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_
tools1
skills/ops/memory-daily-consolidator
# memory-daily-consolidator **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 45 3 * * * ## Purpose Consolidates agent outputs and life events into episodic_memory. Runs after life events extraction. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_
tools1
skills/ops/ocf-agent-report
# ocf-agent-report **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 05 0 * * * ## Purpose Reports CMA/PAR/SEOS agent lifecycle status to Slack #orion-ops. Checks for lifecycle events, detects stuck runs, calculates success rates. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_i
tools1
skills/ops/pii-redaction
# pii-redaction **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 0 2 * * 0 ## Purpose Weekly soft-deletion of expired PII from ClickHouse and Postgres. Nullifies PII fields while preserving row structure for analytics. Targets: ocf_actors.display_name (30d), audit_log IP/UA (90d). ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write inclu
tools1
skills/ops/program-gate-check-namespace-isolation
# program-gate-check-namespace-isolation **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** manual ## Purpose Validates that namespace isolation test passes (Phase 1 gate). ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | From Langfuse or run_i
tools1
skills/ops/validate-cma
# validate-cma **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 47 23 * * * ## Purpose Validates CMA agent output quality, performance, and stability. Triggers auto-rollback if critical failures detected. Checks progression criteria after validation. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-----
tools1
skills/ops/video-storyboard-demo
# video-storyboard-demo **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 0 5 * * * ## Purpose Example automation that calls the Atlas video-generator. This task exits immediately unless VIDEO_AUTORUN_ENABLED=true so we do not consume GPU time without an explicit switch. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Fie
tools1
skills/tbm/tbm-newsletter
# tbm-newsletter **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 0 10 * * 0 ## Purpose Weekly newsletter generation with Slack approval workflow. 1. Fetches top-performing readings from ClickHouse 2. Generates HTML newsletter 3. Saves draft to Supabase (status: pending_approval) 4. Sends Slack notification to #tbm-news with approve instructions Approval via /approve command triggers send through Vercel API. ## Memory Contract ### Access Mode **`service_only`** - All
tools1
skills/tbm/tbm-supabase-sync
# tbm-supabase-sync **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 12 * * * * ## Purpose Push the latest TikTok readings from ClickHouse into the Supabase tiktok_posts table so tarotbymarie.com shows new content. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated
tools1
skills/ops/langfuse-memory-sync
# langfuse-memory-sync **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 34 0 * * * ## Purpose Reads traces from Langfuse, extracts key insights from agent executions, and writes memories to ClickHouse for long-term retention and RAG retrieval. Creates a feedback loop: traces -> memories -> better agent context. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill #
tools1
skills/ops/langfuse-auto-heal
# langfuse-auto-heal **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** */10 * * * * ## Purpose Every 10 minutes, verifies Langfuse health on edge-01 and restarts langfuse/langfuse-worker via Doppler if unhealthy. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUI
tools1
skills/oia/gtm-omega-daily
# gtm-omega-daily **Lane:** oia **Version:** 0.1.0 **Owner:** system **Schedule:** 3 6 * * * ## Purpose Daily viability pulse check for Cosmocrat's GTM health. Runs Layer 0 (Context Lock), Layer 1 (Market Viability), and Block P (Drift Detection). Compares against previous run and alerts on significant drift. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Ever
tools1
skills/ops/agent-cma
# agent-cma **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 42 23 * * * ## Purpose Extracts durable memories (preferences, decisions, blockers, goals) from daily conversation exports using Ollama/OpenAI. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `
tools1
skills/ops/context-post-verify
# Context Post-Verify Skill Verifies model response by: - Checking if retrieved context was actually used - Validating against known invariants - Flagging potential hallucinations - Emitting machine verdict (pass/fail/warn) ## Phase **post_verify** - Runs after model response ## Priority 10 (runs first in post-verify phase) ## Mutations None (read-only verification) ## Verification Checks 1. **Context Usage**: Checks if response mentions concepts from retrieved items 2. **Invariant Vali
testing1
skills/ops/coo-action-engine-execute
# coo-action-engine-execute **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** */5 * * * * ## Purpose Executes due tasks from the COO action queue (dry-run by default). Set ACTION_ENGINE_ENABLED=true to actually execute tasks. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` |
tools1
skills/ops/generate-daily-exec-report
# generate-daily-exec-report ## Purpose Generates the **Executive Daily Report** - a comprehensive daily briefing that synthesizes: - Raw conversations from ChatGPT, Claude, Gemini, Cursor - Extracted memories (CMA output) - Business insights (PAR output) - SEO recommendations (SEOS output) - Operational telemetry ## What It Does 1. **Reads** exported conversations for the date (`/cosmocrat/inbox/conversations/{date}.json`) 2. **Queries** ClickHouse for memories, insights, and telemetry 3. *
tools1
skills/ops/governance-canary-auto-merge
# governance-canary-auto-merge **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 35 5,17 * * * ## Purpose Checks canary validation results and auto-merges governance config changes if valid. Rolls back if canary fails. Runs after canary-pre-flight. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|
tools1
skills/ops/memory-spool-drain
# memory-spool-drain **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** */5 * * * * ## Purpose Drains the JSONL spool files to ClickHouse when they exist. This is the "parachute" for events that couldn't be written directly. Retries with backoff, quarantines poison files after max attempts. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every wri
tools1
skills/ops/morning-coo-briefing
# morning-coo-briefing **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 12 9 * * * ## Purpose Synthesizes Executive Daily Report, memory health, and agent outputs into a concise COO-facing morning briefing. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | |
tools1
skills/ops/ops-memory-scout
# ops-memory-scout **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 2 9 * * * ## Purpose Scans memory systems for anomalies, stale data, and optimization opportunities. Generates daily memory health report. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | |
tools1
skills/ops/program-gate-check-lifecycle-events
# program-gate-check-lifecycle-events **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** manual ## Purpose Validates that lifecycle events are present in ClickHouse (Phase 1 gate). ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | From Langfuse
tools1
skills/ops/program-runner-nightly
# program-runner-nightly **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 0 2 * * * ## Purpose Executes program phases in dependency order, checks gates, and emits telemetry to ClickHouse. Runs Phase 1 work items for cosmocrat-mvp. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run
tools1
skills/ops/program-work-item-run
# program-work-item-run **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** manual ## Purpose Executes a single work item standalone (triggered via MCP or Control Panel). Validates dependencies before execution. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID |
tools1
skills/tbm/tbm-rag-index
# tbm-rag-index **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 35 4 * * * ## Purpose Daily refresh of RAG embeddings for new transcripts. Updates the vector index used by the Marie-voice chatbot. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_i
tools1
skills/tbm/tiktok-health-monitor
# tiktok-health-monitor **Lane:** tbm **Version:** 0.1.0 **Owner:** system **Schedule:** 24 * * * * ## Purpose Hourly health check for TikTok Auto-Gen pipeline. Monitors video generation, QC pass rates, upload success. Alerts to Slack #tiktok-autogen on issues. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--
tools1
skills/ops/validate-par
# validate-par **Lane:** ops **Version:** 0.1.0 **Owner:** system **Schedule:** 53 23 * * * ## Purpose Validates PAR agent output quality and performance. ## Memory Contract ### Access Mode **`service_only`** - All writes go through MemoryClient. No direct ClickHouse access. ### Writes None - read-only skill ## Provenance Every write includes: | Field | Source | |-------|--------| | `run_id` | Generated UUID | | `trace_id` | From Langfuse or run_id | | `source_ref` | `validate-p
tools1