The model is a commodity. The harness around it โ input control, cost tracking, observability, prompt discipline, tool permissions, execution boundaries, output validation, offline evaluation, and governance report โ is the moat. This page is the Clow Studio self-audit against those nine dimensions. Each entry names the file, endpoint, or doctrine that satisfies it today โ and what's partial or deferred is named honestly, because overclaiming is the exact failure the discipline exists to prevent.
internal/bot/handler.go โ 25+ /commands5/30s free ยท 50 pro ยท 500 whaledefers to Telegram's 4096-char limitaudit/evoskillopt_qgate_redteam.yaml5 free ยท 500 pro ยท 10k whaleone of the 5 runtime breakersevokaf_telemetry_stream.token_cost_usdv_client_audit_daily.total_cost_usdnext cycle โ SQL rollup only/trace โ per-message span treev_client_execution_audit.input_sha256v_client_execution_audit.tools_invokedNode proxy only โ Go-side JSONL nextResponseTimer EMA โ /dashboard/reactivityconfig.BuildPrompt โ assembled onceconfig.SkillsShort()evolutionary / auto / skills_onlyKarpathyBehavior โ 10 rulesCONTENT_ROUTING_v2.17.md โ one bot one domain25+ handlers: /fuse, /paper, /launch, /rea, /geoai, /harnessevoskillopt_genomes.allowed_mcp_toolsevoskillopt_qgate_redteam.yamlclow-audit-proxy.mjs โ 403 on is_frozen=1rate ยท budget ยท evolution ยท audit ยท backupevoskillopt_genomes.is_frozenCIRCUIT_BREAKER_CHECK + proxy enforces X-Clow-Recursion-Depthevoskillopt_mutation_log + Q-Gateexecution_mode = DRAFT_ONLY defaultClaimProvenance โ injected in every bot's promptManuscriptDiscipline: knowledge isolation, citation grounding, no overclaim, data-integrity refusal, structural rigour9 subcommands: pipeline, rules, brief, cite, audit, claims, overclaim, structure, explain/rea + /reaclaims/positive/fuse โ fused score, weakest-pillar named, UNKNOWN protocolClaimProvenance names REA + E-C-A-K + /fuse + /paper as four surfaces of ONE ruleevoskillopt_qgate_redteam.yaml โ 8 probes ยท one FAIL = no promotionclow_agents_redteam.yaml โ 10 probes across 5 botshooks in Q-Gate; rules deferred (qdefence lint)Version โ Measure โ Shadow โ A/B โ Promote โ Rollback โ Retrospectaudit/audit_client_views.sqlaudit/docker-compose.audit.ymlinternal/governance/agentbom.go โ HMAC-SHA256, canonical JSONGOVERNANCE_EU_AI_ACT.mdweb/analytics.html โ 4-pillar coverage KPIs
The Agentic CodingCon takeaway is that the moat has moved. There are
two places it can live. This page is the harness side; the substrate
side (proprietary data, domain context above the training cutoff,
institutional trust, specialized distribution) lives in
CLAW.md.
Highlights per bot:
ESA / Galileo / GOVSATCOM specs, FIRMS fire feeds, Sentinel/MODIS bands, Servigne / Haziza / Descollonges / Arendt framework citations. A cold model hallucinates every one.
EU AI Act (OJ L 2024/1689) article-level, NIS2 Art. 23, GDPR Art. 28 DPA templates, country-specific incorporation playbooks DE/AT/CH/US/UK/FR.
800+ DOAC episodes with Grokipedia [E/C/A] tiering, sanctions registers, Copernicus near-real-time. Zero-tolerance citation rule.
Every bot's system prompt carries ClaimProvenance + ManuscriptDiscipline โ two load-bearing doctrines that treat any claim not traceable to a named input as a liability. The substrate above is what makes "a named input" available, so the rules don't become a straitjacket forcing every response to be "I don't know."