sicambria
Community@sicambria
sicambria provides a gated orchestration harness for disciplined software delivery: phase-gated planning, adversarial review, multi-agent backlog draining, and self-verified deploys.
Agent Skills by sicambria
Showing 11 vetted skills indexed across 1 GitHub repositories.
sicambrianext
Orchestrates one unit of work through enforced plan, analyze, approval, execute, and review phase gates.
sicambriadecide
Writes dated decision records and logs decision events to an observability log.
sicambriared-team
Attacks code changes with adversarial inputs to expose defects before shipping.
sicambriaagent-protocol
Orchestrates multi-agent waves to triage, delegate, and serially integrate backlog work.
sicambriadeploy
Runs verified deploys with healthchecks, event logging, and automatic rollback on failure.
sicambriadebugging
Diagnose software defects through hypothesis-driven root-cause analysis before applying fixes.
sicambriatodo
Enumerates and triages every backlog surface, then chains actionable items into agent-protocol.
sicambriaretro
Synthesizes dated retros from the observability event log into committed archive files.
sicambriaanalyze
Detects contradictions, coverage gaps, and scope drift across plan, spec, and code.
sicambriacode-quality
Assess code changes for reuse, minimality, and correct abstraction altitude.
sicambriaplanmax
Scores implementation plans against a rubric and iterates them before coding.
Frequently Asked Questions About sicambria
FAQPage SchemaWhat tasks can I accomplish with sicambria's skills?▼
You can run gated plan-to-execution spines (/next), drain backlogs with multi-agent waves (/agent-protocol), adversarially attack changes (/red-team), check plan-vs-code drift (/analyze), score and iterate plans (/planmax), debug root causes, deploy with auto-rollback, and log auditable decisions and retros.
Who are these skills designed for?▼
Software engineers and technical leads running disciplined, agent-assisted development who need enforced phase gates, adversarial review before shipping, auditable decision records, and serialized integration of parallel subagent work rather than ad-hoc guess-and-check changes.
How do the skills run in practice?▼
Skills are invoked via slash-command triggers like /next, /deploy, or /red-team. The /next spine is enforced by .harness/scripts/next-phase-gate.mjs, which blocks out-of-order transitions; events are logged to .harness/state/events.jsonl and artifacts archived under .harness/archive/.
What prerequisites or dependencies do these skills require?▼
A repository with the .harness directory structure, including scripts/next-phase-gate.mjs, state/events.jsonl, and archive folders. Deploys require a config.deploy entry with a deploy command; the deploy skill abstains rather than silently passing if none is configured.
How does the multi-agent backlog protocol work?▼
The agent-protocol skill triages and clusters backlog items, spawns one wave of agents that each plan and implement in a git worktree, then serially integrates their branches through the orchestrator with adversarial verification, running to exhaustion and documenting skips.