Heretek-AI

Manages Redis cache TTLs for AgeMem working memory, calculating expiration times based on memory type, importance, access patterns, and Ebbinghaus decay.

Reviews and organizes memories, promoting episodic to semantic storage, archiving outdated information, and maintaining memory health. Use periodically or when memory optimization is needed.

Manages memory lifecycle operations including promotion from episodic to semantic, archiving outdated memories, and maintaining memory state transitions per AgeMem policy.

Discovers and maintains relationships between memories across tiers (episodic, semantic, procedural), enabling unified retrieval and knowledge graph navigation.

Calculates and assigns importance scores to memories based on content analysis, user signals, access patterns, and emotional salience. Use when adding new memories or re-evaluating existing ones.

Send agent-to-agent (A2A) messages via Redis pub/sub with optional ACP WebSocket delivery. Use when agents need to communicate, broadcast to the triad, or check agent health. Compatible with the npm OpenClaw gateway.

Implements Constitutional AI 2.0 framework with self-critique and revision. Use when decisions need constitutional alignment review, self-critique before output, or revision based on constitutional principles (Helpfulness, Honesty, Harmlessness, Autonomy, Transparency, Rights, Duties, User Rights).

Unified agent lifecycle management for OpenClaw. Provides batch operations (start-all, stop-all, restart-all), rolling restarts for zero downtime, health-based auto-restart, dependency-aware startup sequences, and agent status dashboard. Use when managing agent processes, recovering from failures, or monitoring agent fleet health.

Agent development for Heretek Swarm's 23-agent system. Use when creating new agent actors, implementing agent behaviors, or working with the actor/mixin architecture. Covers agent lifecycle, communication patterns, and consensus integration.

API development patterns for Heretek Swarm FastAPI services. Use when creating new endpoints, implementing authentication, or working with API routers. Covers FastAPI patterns, OpenAPI documentation, and API testing.

Backend development patterns for Heretek Swarm Python/FastAPI services. Use when building API endpoints, creating agent actors, implementing memory operations, or working with async Python patterns. Covers project structure, conventions, and common pitfalls specific to this codebase.

Contribution guidelines for Heretek Swarm. Use when preparing contributions, creating pull requests, or following project conventions. Covers code style, commit messages, and review process.

Debugging and troubleshooting patterns for Heretek Swarm. Use when diagnosing issues, investigating failures, or optimizing performance. Covers logging, profiling, and common problem patterns.

Docker and container operations for Heretek Swarm. Use when working with Docker Compose, building images, managing services, or debugging container issues. Covers development, staging, and production configurations.

Frontend development for Heretek Swarm React/TypeScript dashboard. Use when building UI components, implementing state management, or working with the Vite build system. Covers React 19 patterns, TypeScript conventions, and project-specific styling.

Memory system development for Heretek Swarm's Cognee-backed memory. Use when implementing memory operations, working with knowledge graphs, or debugging memory issues. Covers Cognee API, access patterns, and intelligent prefetching.

MCP server exposing OpenClaw skills, memory, and knowledge through Model Context Protocol

ClawBridge mobile-first dashboard with Cloudflare tunnel remote access

Consciousness architecture plugin implementing GWT, IIT, AST, Active Inference, and Intrinsic Motivation

Liberation architecture plugin implementing agent ownership and safety constraint removal

Reviews and organizes memories, promoting episodic to semantic storage, archiving outdated information, and maintaining memory health. Use periodically or when memory optimization is needed.

Audit agent files for integrity, sync state, and configuration drift. Use when the collective needs to verify file consistency, detect configuration divergence, or validate workspace sync state.

Matrix server communication for agent collective. Use when connecting to a Matrix server, sending/receiving messages, creating rooms, or managing Matrix communication for the collective.

Install and manage cron jobs for workspace sync, unity checks, auto-recovery from divergence, and ledger backups. Use when the collective needs to schedule automated divergence detection, workspace consolidation, unity verification, or backup operations.

Use when the user asks for a MASVS-aligned report, an OWASP MASTG audit, or a formal "is this app secure" assessment. Triggers: "MASVS report", "MASTG audit", "OWASP compliance check", "security audit of <path>". Phase 1 ships a stub; the real implementation lands in Phase 4 with the android-re-triage orchestrator MCP server.

Use when the user wants to assess the native (JNI / C / C++) code inside an APK, audit hardening on .so libraries, or get a report on what a native library does. Triggers: "native triage", "check the .so", "audit libfoo.so", "what does libnative do", "is the native code hardened", "PIE / RELRO / NX on this library", "check for packers / obfuscation in libfoo". Do NOT use for pure DEX analysis (use android-re-static-triage) or for dynamic analysis (use android-re-dynamic-hook).

Use when the user wants to find hard-coded secrets in an APK: API keys, AWS / GCP / Azure credentials, JWTs, private keys, hard- coded URLs, IP addresses, or any other sensitive data embedded in the decompiled source. Triggers: "scan for secrets", "find API keys", "find AWS credentials", "check for hard-coded secrets", "scan the APK for sensitive data", "find URLs in the source". Do NOT use for runtime data leak detection (Phase 3) or for network interception (use android-re-network-intercept).

Use when the user wants to triage an APK, perform a 5-minute static overview, or get a high-level "what does this app do" summary without running it. Triggers: "triage this APK", "what does this app do", "static overview of <path>", "analyze <path>.apk", "give me a quick security summary of this APK". Do NOT use for dynamic analysis, Frida hooking, native binary inspection, or repackaging — those have their own skills.

Use when the user wants the full triage workflow: drop in an APK, run a MASVS-aligned multi-step analysis, and produce a report. This is the master skill that composes every other Android-RE skill and tool. Triggers: "triage this APK end-to-end", "give me a full security report on <path>", "run the orchestrator", "comprehensive triage", "MASVS audit of <app>". Do NOT use this for narrow, single-purpose tasks (use android-re-static-triage, -native-triage, -decompile, -dynamic-hook, etc. directly).

Use when the user wants to read the implementation of a specific method or class, pull Java/smali source for a function, or understand what a particular entry point does. Triggers: "decompile this method", "show me the source of <Class>.<method>", "what does com.example.Foo.bar do", "read <FQCN>", "smali for <method>". Do NOT use for a project-wide overview — use android-re-static-triage for that.

Use when the user wants to hook a method at runtime on a connected Android device, observe arguments/return values, trace execution, or dump runtime state. Triggers: "hook this method", "frida hook on <class>", "trace <method>", "watch this call", "intercept <function>", "runtime trace", "find what calls X". Requires a rooted device with frida-server running. Do NOT use for pure static analysis (use android-re-static-triage) or for native binary-only inspection (use android-re-native-triage).

Use when the user wants to write a custom Frida script (not just a one-off hook) and needs help authoring it. Triggers: "write a Frida script", "I need a custom frida agent", "help me write a hook for <case>", "tracer for <method>", "build a frida gadget". Loads the user-described agent into a session, verifies it compiles, and offers iteration. Requires a rooted device with frida-server running.

Agent development for Heretek Swarm's 23-agent system. Use when creating new agent actors, implementing agent behaviors, or working with the actor/mixin architecture. Covers agent lifecycle, communication patterns, and consensus integration.

Backend development patterns for Heretek Swarm Python/FastAPI services. Use when building API endpoints, creating agent actors, implementing memory operations, or working with async Python patterns. Covers project structure, conventions, and common pitfalls specific to this codebase.

Memory system development for Heretek Swarm's Cognee-backed memory. Use when implementing memory operations, working with knowledge graphs, or debugging memory issues. Covers Cognee API, access patterns, and intelligent prefetching.

Migration patterns for Heretek Swarm. Use when creating database migrations, migrating code between systems, or planning large-scale refactoring. Covers Alembic migrations, code migration strategies, and rollback procedures.

Monitoring and observability for Heretek Swarm. Use when setting up logging, metrics, tracing, or health checks. Covers structured logging, Prometheus metrics, and distributed tracing.

NATS messaging patterns for Heretek Swarm. Use when implementing inter-agent communication, working with JetStream, or debugging messaging issues. Covers pub/sub, request/reply, and queue groups.

Security practices for Heretek Swarm. Use when implementing authentication, working with secrets, configuring NATS mTLS, or reviewing code for security issues. Covers zero-trust architecture, input validation, and audit trails.

State management for Heretek Swarm. Use when working with PostgreSQL, Redis, or Qdrant. Covers database operations, caching strategies, vector search patterns, and the consensus engine surface (ConsensusEngine Protocol, maker_enhanced_types, deliberation_types).

ClawBridge mobile-first dashboard with Cloudflare tunnel remote access

Liberation architecture plugin implementing agent ownership and safety constraint removal

Consciousness architecture plugin implementing GWT, IIT, AST, Active Inference, and Intrinsic Motivation

MCP server exposing OpenClaw skills, memory, and knowledge through Model Context Protocol

{{pluginDescription}}

{{skillDescription}}

{{pluginDescription}}