🤖

Agents

Personas with autonomy policies — recommend, approve, low-risk-auto, full-auto.

39

vetted (of 57)

🔍 Tags across Agents

Click any tag to filter the inventory.

🔢Order = frequency (matches) · 📏Size = recency (newer → bigger) · 🎨Color = avg rating · ↗Tilt = vetted ratio (upright → vetted)

colaberry openclaw agent security detective-agent frontend agents validation scoring orchestration multi-platform polling anomaly-detection gpt-4o conversion autonomy attribution follow-up classification quality-gate multi-agent pipeline llm personalization planner memory secret-scanning static-analysis regex-rules ticket-emitter learning feedback-loop optimization aggregation signal-ingestion social-listening runtime-monitoring scraping-detection abuse-detection llm-generation round-robin-queue draft-pipeline category-balancing sast vulnerability-scanning owasp testing playwright-poster anti-detection daily-rate-cap

Filter: All ✓ Colaberry vetted only 📁 source: execution/ops_platform/agent_registry.py 🌐 Ingest URL + Submit manually

📋 Agents inventory

39 entries

secret-detection-agent Colaberry vetted ✓ Colaberry Inc approved

Filesystem-walking detective that scans backend/src and frontend/src for hard-coded secrets using a curated regex library (AWS keys, private keys, DB connection strings, JWTs, OpenAI sk- keys, Supabas

securitysecret-scanningdetective-agentstatic-analysisregex-rulesticket-emittercolaberry

ali@colaberry.com

openclaw-learning-optimization-agent Colaberry vetted ✓ Colaberry Inc approved

Engagement-feedback learning engine. Reads all posted OpenclawResponses with engagement_metrics, groups by tone/timing/topic, computes average engagement (upvotes + 2×replies + 3×clicks), writes findi

openclawlearningfeedback-loopoptimizationaggregationcolaberry

ali@colaberry.com

openclaw-market-signal-agent Colaberry vetted ✓ Colaberry Inc approved

Multi-platform signal-scanning agent that polls 11+ public APIs (Reddit, HackerNews, Dev.to, Hashnode, Discourse, Twitter, Bluesky, YouTube, ProductHunt, Facebook Groups, LinkedIn comments) for AI-rel

openclawsignal-ingestionmulti-platformsocial-listeningpollingcolaberry

ali@colaberry.com

runtime-threat-monitor-agent Colaberry vetted ✓ Colaberry Inc approved

Real-time abuse detection on production telemetry. In a rolling 5-minute window, groups PageEvent rows by visitor_id (flags visitors with >30 events as suspicious, >100 as critical scraping) and group

securityruntime-monitoringscraping-detectionabuse-detectionanomaly-detectioncolaberry

ali@colaberry.com

platform-content-response-agent Colaberry vetted ✓ Colaberry Inc approved

Queue-consumer agent that pulls pending generate_reply / generate_post tasks via round-robin across content categories (preventing single-category starvation), calls GPT-4o with a per-category system

llm-generationround-robin-queuedraft-pipelinecategory-balancinggpt-4ocolaberry

ali@colaberry.com

code-security-audit-agent Colaberry vetted ✓ Colaberry Inc approved

Static AST-light vulnerability scanner that walks backend TypeScript and flags SQL injection (template-literal queries, string-concat SQL), XSS (innerHTML, dangerouslySetInnerHTML, document.write), co

securitysastvulnerability-scanningowaspdetective-agentcolaberry

ali@colaberry.com

conversion-testing-agent Colaberry vetted ✓ Colaberry Inc approved

Agent that audits pages for conversion optimization best practices. Integrates with the agent orchestrator to provide recommendations on form optimization and CTA placement.

frontendagentsconversiontestingcolaberry

ali@colaberry.com

rate-limited-browser-poster-agent Colaberry vetted ✓ Colaberry Inc approved

Browser-automation worker that posts approved drafts to a target platform with built-in anti-detection: randomized human-like delays between actions, hard daily post cap enforced via timezone-aware SQ

playwright-posteranti-detectiondaily-rate-captimezone-awareapproved-only-gatecolaberry

ali@colaberry.com

cory-strategic-agent Colaberry vetted ✓ Colaberry Inc approved

On-demand strategic planning agent for Cory briefing engine. Translates natural language commands into structured tickets with sub-tasks. Operates under command intent: plan_curriculum, fix_platform,

strategic-planningagentautonomycolaberry

ali@colaberry.com

dependency-security-agent Colaberry vetted ✓ Colaberry Inc approved

Runs `npm audit --json` against each package.json under the repo (backend, frontend, etc.), aggregates critical/high/moderate/low counts, captures up to 15 advisories per project with module name + ad

securitysupply-chaincvenpm-auditdetective-agentcolaberry

ali@colaberry.com

openclaw-infra-monitor-agent Colaberry vetted ✓ Colaberry Inc approved

Browser-session health monitor + self-healer. Watches every OpenclawSession row, detects captcha blocks (auto-pauses all platform tasks), counts recent errors (>5 in 1h triggers session restart if aut

openclawinfra-monitorself-healcaptcha-detectionsession-healthcolaberry

ali@colaberry.com

proactive-outreach-flagger-agent Colaberry vetted ✓ Colaberry Inc approved

Scheduled flagger that watches the IntentScore table for visitors crossing a configurable threshold (default score >= 60) and sets a proactive_chat_pending flag on the visitor record so the frontend c

proactive-outreachintent-drivenui-handoffflag-basedcolaberry

ali@colaberry.com

openclaw-conversion-detection-agent Colaberry vetted ✓ Colaberry Inc approved

Stage-5-to-stage-6 conversion detector. Finds active conversations at stage 5 (interest expressed), checks ResponseQueue for a call/link offer, looks for positive response in the engagement chain, adv

openclawconversionpipelineattributionstage-advancementcolaberry

ali@colaberry.com

ux-ui-validation-agent Colaberry vetted ✓ Colaberry Inc approved

Read-only agent that validates pages against UX/UI design rules using a blueprint validator service. Produces detailed reports with per-check confidence levels.

frontendagentsvalidationux-uicolaberry

ali@colaberry.com

notification-question-classifier-agent Colaberry vetted ✓ Colaberry Inc approved

Heuristic classifier that scans community notification feeds and DM inboxes for messages worth replying to. Uses two pattern lists - QUESTION_INDICATORS (question marks plus 'how do', 'what is', 'can

question-detectionreaction-filterai-loop-preventiondeterministic-classifierinbox-triagecolaberry

ali@colaberry.com

ai-safety-monitor-agent Colaberry vetted ✓ Colaberry Inc approved

Scans the last 5 minutes of inbound chat messages (role=visitor|user) for prompt-injection and jailbreak patterns across 10 curated categories: system-prompt override, role impersonation, system-messa

ai-safetyprompt-injectionjailbreak-detectionchat-monitoringdetective-agentcolaberry

ali@colaberry.com

conversation-task-monitor-agent Colaberry vetted ✓ Colaberry Inc approved

Regex-pattern scanner that detects actionable requests inside active live-chat conversations (callback, document send, email me, text me, schedule a call) and materializes them as pending action recor

task-detectionregex-routingchat-to-actionidempotentcolaberry

ali@colaberry.com

agent-behavior-monitor-agent Colaberry vetted ✓ Colaberry Inc approved

Meta-agent that watches the AI agent fleet itself. Detects (1) stuck agents (status=running for >15min), (2) error spikes (>=5 failed runs in last hour), (3) duration anomalies (last run >3x rolling-a

agent-opsmeta-monitoringanomaly-detectionfleet-healthstall-detectioncolaberry

ali@colaberry.com

access-control-guardian-agent Colaberry vetted ✓ Colaberry Inc approved

Walks backend Express routes and detects access-control gaps: admin-prefixed routes missing requireAdmin middleware, mutating routes (POST/PUT/PATCH/DELETE) missing any auth guard, and missing Helmet

securityauthzrbac-auditexpressmiddleware-auditcolaberry

ali@colaberry.com

openclaw-follow-up-agent Colaberry vetted ✓ Colaberry Inc approved

Stale-conversation revival agent. Targets conversations with stall_detected_at IS NOT NULL at stage >= 2 and priority_tier in (hot, warm). Generates stage-aware follow-up drafts via GPT-4o, validates

openclawfollow-upstale-revivalstage-awaredraft-onlycolaberry

ali@colaberry.com

openclaw-browser-worker-agent Colaberry vetted ✓ Colaberry Inc approved

Multi-strategy posting worker that processes approved OpenclawResponses through a layered fallback: API posting → Playwright browser automation → retry with backoff → human queue. Enforces HUMAN_EXECU

openclawposting-workerfallback-chainbrowser-automationrate-limitingcolaberry

ali@colaberry.com

high-intent-corporate-lead-classifier-agent Colaberry vetted ✓ Colaberry Inc approved

Scheduled scanner that promotes ambiguous visitors to 'enterprise' classification when corporate-email-domain + engagement-depth + interest-signal criteria all line up. Uses an explicit FREE_EMAIL_PRO

lead-scoringclassificationenterprise-detectioncorporate-emailcolaberry

ali@colaberry.com

openclaw-supervisor-agent Colaberry vetted ✓ Colaberry Inc approved

Task-queue supervisor that watches OpenclawTask + OpenclawSession state. Cancels tasks stuck in 'running' beyond threshold (auto-retry up to max_retries, then fail), enforces per-platform rate budgets

openclawsupervisortask-queuestuck-detectionrate-budgetscolaberry

ali@colaberry.com

openclaw-response-orchestrator-agent Colaberry vetted ✓ Colaberry Inc approved

Conversation-aware reply-generation agent. For each new EngagementEvent above intent threshold, rebuilds full conversation context (original signal + our prior response + their reply chain), detects c

openclaworchestratorconversation-stagellmfollow-upcolaberry

ali@colaberry.com

strategy-architect-agent Colaberry vetted ✓ Colaberry Inc approved

Parameterized strategic-planning agent that runs a per-department cycle: (1) evaluates department health + innovation scores, (2) loads existing planned/active initiatives to suppress duplicates, (3)

strategyllm-planninginitiative-creationparameterized-agentgpt-4ohybrid-rule-llmcolaberry

ali@colaberry.com

openclaw-content-response-agent Colaberry vetted ✓ Colaberry Inc approved

Generates platform-appropriate insight-only social replies for queued signals using GPT-4o with template fallback. Per-response unique marketing tag (oc-{platform}-{8hex}) generates a tracked URL for

agentcontent-generationmulti-platformattributionoutboundcolaberry

ali@colaberry.com

community-signal-detection-agent Colaberry vetted ✓ Colaberry Inc approved

Playwright-driven scanner that crawls a closed online community (Skool/Discord/Circle), navigates per-category, extracts posts via DOM-stable link detection (resilient to styled-components hash churn)

community-listeningplaywrightsignal-detectionqueue-producersocial-monitoringcolaberry

ali@colaberry.com

skool-quality-gate-agent Colaberry vetted ✓ Colaberry Inc approved

Pure deterministic validation of draft Skool responses. No LLM calls. Checks length, banned words, URL compliance, content quality, duplicate detection, sign-off presence, urgency language, case-study

skoolquality-gategovernancedeterministiccolaberry

ali@colaberry.com

security-director-agent Colaberry vetted ✓ Colaberry Inc approved

Coordinator (not detective) that runs every 10 minutes to roll up the security department: gathers recent DepartmentEvents (threats, alerts, scans), counts open and critical security tickets, audits s

securitycoordinator-agentfleet-orchestrationrollupself-healing-ticketscolaberry

ali@colaberry.com

marketing-strategy-agent Colaberry vetted ✓ Colaberry Inc approved

Specialized agent that evaluates marketing strategy alignment on pages. Part of the agent orchestration system and contributes validation results to cross-project marketing blueprint compliance.

frontendagentsmarketingvalidationcolaberry

ali@colaberry.com

agent-orchestrator Colaberry vetted ✓ Colaberry Inc approved

Multi-agent orchestration framework that coordinates specialized agents (UX/UI, marketing, conversion testing) and synthesizes their findings into a unified report with confidence scoring and blueprin

frontendagentsvalidationorchestrationcolaberry

ali@colaberry.com

admissions-intent-detection-agent Colaberry vetted ✓ Colaberry Inc approved

Applies admissions-specific intent thresholds to IntentScore records and updates recommended_next_action on AdmissionsMemory. Runs every 10 minutes. Thresholds: enrollment_ready (80), high_intent (60)

admissionsintent-detectionscoringroutingcolaberry

ali@colaberry.com

openclaw-authority-content-agent Colaberry vetted ✓ Colaberry Inc approved

Scheduled (daily 8am UTC) authority/thought-leadership content generator. Pulls last 48h of high-signal social activity, groups by topic tag, picks top trending topics, queries OpenclawLearning for be

agentscheduledauthoritylong-formsignal-synthesiscolaberry

ali@colaberry.com

openclaw-conversation-detection-agent Colaberry vetted ✓ Colaberry Inc approved

Deterministic scoring + queueing agent that takes 'discovered' OpenclawSignal rows and assigns three scores (relevance, engagement, risk). Above relevance + engagement thresholds AND below risk thresh

openclawscoringtriagequeue-gatingcolaberry

ali@colaberry.com

agent-attribution-classifier-llm Colaberry vetted ✓ Colaberry Inc approved

Autonomous classifier mapping backend services to business capabilities. Runs after each brownfield discovery scan. For each (capability, candidate agent) pair, calls gpt-4o-mini to rate confidence. C

classificationagentautonomycolaberry

ali@colaberry.com

openclaw-engagement-monitor-agent Colaberry vetted ✓ Colaberry Inc approved

Engagement polling agent that scans every posted OpenclawResponse + AuthorityContent URL for new comments/replies on the source platform. Normalizes engagement events, dedupes by source_url+response_i

openclawengagement-trackingpollingmonitorcolaberry

ali@colaberry.com

openclaw-quality-gate-agent Colaberry vetted ✓ Colaberry Inc approved

AI-powered review agent that gates draft responses before posting. Deterministic scoring 0-100 based on length (120-2000 chars), banned phrase regex (DM-bait, promotional), link density (max 2), em-da

agentquality-gatedeterministic-scoringregeneration-loopcolaberry

ali@colaberry.com

conversation-goal-planner-agent Colaberry vetted ✓ Colaberry Inc approved

Deterministic planner that selects a conversation goal and messaging strategy for any visitor based on stored visitor_type, intent level, prior conversation count, and known interests. Returns one of:

conversation-planningpersonalizationgoal-selectiondeterministic-routingcolaberry

ali@colaberry.com

vip-recognition-agent Colaberry vetted ✓ Colaberry Inc approved

Event-driven or batch agent that detects when a visitor's linked lead matches a VIP/CEO pattern, promotes the visitor classification accordingly, and emits a recognition event so downstream agents can

vip-detectionabmidentity-classificationevent-drivencolaberry

ali@colaberry.com