Agent Building skills

Browse reusable Agent Skills, each with a clear purpose and practical guidance.

harness-design

Design and build multi-agent harness architectures for long-running AI application development. GAN-inspired Generator-Evaluator pattern, Sprint Contract negotiation, context management, quality criteria calibration. Based on Anthropic Engineering patterns. Use when: "build a harness", "multi-agent architecture", "agent orchestration", "generator-evaluator", "long-running app", "harness design", "agent pipeline", "quality evaluation loop", "sprint contract", "build app with agents", "Claude Agent SDK architecture", or when building complex full-stack apps that need planning → generation → evaluation cycles. Also use when discussing context degradation, self-evaluation bias, or assumption testing in AI workflows. Do NOT use to stress-test or critique an already-written plan document; use plan-swarm-review for that (this skill designs the harness, it does not review plans).

135 repo starsObserved in 1 repos
Agent Building
GitHub

hephaestus-network

Use when the user types /hep-network, mentions @Hephaestus, or asks Agentlas to staff a task from registered Local, owner Cloud, and public Hub agents or teams. The active host LLM is the temporary orchestrator.

135 repo starsObserved in 1 repos
Agent Building
GitHub

hephaestus-storm

Use when the user types /hep-storm, mentions @Hephaestus storm, or asks to force-robustly drive a loop-worthy goal (apps, sites, agents, automations, debugging, multi-step research, data/report generation) to a verified finish. Stormbreaker routes the goal to real Agentlas specialists, materializes a dependency-ordered pipeline fabric, and runs a verifier-first loop that does not stall, run away, or claim false success. Trivial questions are answered directly, not stormed.

135 repo starsObserved in 1 repos
Agent Building
GitHub

llm-runtime-architecture

Use when designing how one canonical agent core runs across Codex, Claude Code, Gemini CLI, Cursor, and AGENTS.md-compatible tools.

135 repo starsObserved in 1 repos
Agent Building
GitHub

long-horizon-execution

Long-horizon robot task execution workflow for multi-step manipulation tasks. Use this skill when one user goal must be decomposed into ordered subtasks, each subtask needs explicit success checks, retries, or recovery, and every prompt-driven policy rollout should be executed through $monitored-subtask-execution instead of calling raw MCP robot tools directly.

135 repo starsObserved in 1 repos
Agent Building
GitHub

memory-manager

Standardized workflow for discovering, reading, and writing the project's memory file (memory.instructions.md) to persist context across chat sessions.

135 repo starsObserved in 1 repos
Agent Building
GitHub

memory-ticketing

Use when adding Memory Events, Memory Tickets, memory-map.json, vault-references.json, PM Soul memory ownership, or Memory Curator routing to an agent repo.

135 repo starsObserved in 1 repos
Agent Building
GitHub

mode-classification

Use before routing a /meta-agent request to choose single-agent-creator, team-builder, or agentlas-packager from the user's wording and available files.

135 repo starsObserved in 1 repos
Agent Building
GitHub

monitored-subtask-execution

Monitored single-subtask execution workflow. Use the repository MCP service `corobot_mcp_server` to start, monitor, stop, and reset one prompt-driven robot rollout through the sequence `set_evaluate_params`, poll `get_status`, then `stop/reset`. Use this skill whenever a larger workflow needs one repeatable, safe execution unit with timeout handling and deterministic cleanup.

135 repo starsObserved in 1 repos
Agent Building
GitHub

pm-soul

Use when preserving product intent, acceptance criteria, decision memory, open loops, roadmap context, or the product-manager continuity layer of an agent team.

135 repo starsObserved in 1 repos
Agent Building
GitHub