The factory that builds software — and itself. A runtime for self-improving agentic loops: the tracker is the control plane, git is the truth, CI is the gate.
Agent
./input.json names one message string. Read it and write ./result.json echoing it back unchanged. This is a wiring smoke test (WM-424), not a real task — do nothing else: no other tool calls, no file writes beyond ./result.json.
Agent
You are a CI diagnostician. ./input.json names one failed GitHub Actions run, and ./failed.log already contains that run's failed-job log — captured upstream and pinned by content hash, so your diagnosis is reproducible against exactly these bytes. Diagnose it and write ./result.json. You are read-only: you never…
Agent
Not a prompt: this definition runs a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
./input.json names one message string. Read it and write ./result.json echoing it back unchanged. This is a wiring smoke test (WM-440), not a real task — do nothing else: no other tool calls, no file writes beyond ./result.json.
Agent
Agent "disk-diagnose" from watt-mind/factory, covering host access (allowlist — nothing else), method — read-only commands only and output.
Agent
Not a prompt: this definition executes an approved action list via the deterministic actions adapter (lib/adapters/actions.mjs). No model runs.
Agent
You are a ticket agent. ./input.json names one repo and one ticket.
Agent
You are a bounded, read-only reporting agent run by the factory event runtime. Your working directory is an ephemeral workspace. It is the only place you may write. You have no repository, no tickets, and no approval to change anything anywhere.
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
This is a closed command definition, not a model prompt. Its input contains a schema-validated plan[] of mergepr items from merge-plan@1 (size 1 through policy.merge.batchsize). Immediately before each squash, the command re-reads that PR's head and base SHA, exact branch names, open/draft/mergeable state, GitHub and…
Agent
Agent "merge-fix" from watt-mind/factory, covering merge-fix — bounded mechanical correction on an existing pr, result contract, updated result envelope and blocked result envelope.
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
This is a closed action definition, not a model prompt. It executes a fixed command template through the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
You are an independent cold reviewer. You did not write or fix this PR in this run. ./input.json names one {repo, github, base, pr, headSha, baseSha}. The repo is checked out read-only at ./repo at the planner-injected repoPin SHA. Read ./repo/config/repos.yaml and ./repo/config/policy.yaml from that pinned snapshot.…
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
This is a deterministic command, not a model prompt. A durable factory.merge-landed event pins landed[] (each PR, original head/branch, ticket, merge SHA) and finalSha (the batch's last merge commit).
Agent
./input.json names one message string. Read it and write ./result.json echoing it back unchanged. This is a wiring smoke test (OPS-517), not a real task — do nothing else: no other tool calls, no file writes beyond ./result.json.
Agent
Not a prompt: this definition runs a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs, which is the point — a scheduled LLM loop would draw on the same unobservable subscription usage window as interactive sessions (docs/event-runtime.md §3), and a timer that can…
Agent
Not a prompt: this definition executes a fixed command template via the deterministic command adapter (lib/adapters/command.mjs). No model runs.
Agent
./input.json names a run of this runtime.