project-flood
v0.2.0by Slight76
A governed Copilot squad that can create bounded research and worktree swarms.
What project-flood does, in the publisher's words
Project Flood is a governed GitHub Copilot squad that can fan out into temporary research, review, or worktree-isolated build swarms. v0.2 turns the v0.1 operating contract into a portable Agent Plugins 1.0 package plus a repository-owned adapter.
The design stays intentionally bounded:
- one human-facing Flood Squad Lead;
- eight hidden specialists, including a dedicated Flood Integrator;
- maximum three active workers per wave;
- read-only research, architecture, verification, security, and swarm planning;
- separate worktree sessions and exclusive paths for parallel editors;
- deterministic hooks for dangerous commands, protected paths, ownership, and external writes;
Project Flood does not train or modify a model. Its learning is reviewable repository context.
From the project README, punctuation lightly normalized · Open the README on GitHub
- Skills
- 10
- MCP servers
- 0
- Stars
- 0
- License
- n/a
- Repo created
- 2026-08-20
- Last pushed
- 2026-08-21
- Publisher type
- User
- Version
- 0.2.0
Links
Skills · 10
- flood-context-hygiene
- Revalidate, compact, archive, or discard stale Project Flood repository context when memory or orchestration state has become noisy or contradictory.
- flood-dependency-aware-swarm
- Decide whether multi-part work merits a bounded squad or swarm, then define safe waves, dependencies, ownership, gates, and convergence.
- flood-memory-governance
- Verify, promote, supersede, archive, or reject durable Project Flood memory after accepted work, explicit decisions, or revalidation.
- flood-repository-onboarding
- Establish or refresh an evidence-backed Project Flood profile, routes, commands, harness capabilities, and environment setup before substantial work.
- flood-run-feature
- Coordinate a bounded feature or fix from evidence and acceptance criteria through implementation, independent verification, and memory reflection.
- flood-secure-change-review
- Review a security-sensitive change involving trust boundaries, identity, input, data, dependencies, CI/CD, hooks, MCP, or external tools.
- flood-spec-workflow
- Add proportional specification rigor before implementation when product behavior, interfaces, migrations, or cross-team contracts are materially unresolved.
- flood-squad-health
- Audit Project Flood discovery, routing, memory freshness, hooks, manifests, ownership, compatibility, and current blockers without changing files.
- flood-verification-gate
- Independently verify a behavior-changing implementation against acceptance criteria, regression risk, and observable evidence before completion.
- flood-worktree-swarm
- Plan and control parallel implementation in isolated Git worktree sessions with a validated ownership manifest and explicit fan-in.
Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.
Category
Agent Tooling. Assigned by this directory. Agent Plugins 1.0.0 has no category field, so no manifest declares one.
Keywords
copilot · agents · squad · swarm · governance · memory
Related by publisher and keywords
Project-scoped durable memory, recall, Skills, and governed context for AI agents.
Build, deploy, and operate AI agents on AWS. Skills for scaffolding agents with Amazon Bedrock AgentCore (Strands, LangGraph), connecting tools via Gateway and MCP, multi-agent and A2A orchestration, memory, Cedar policies, evaluation, observability, debugging traces and logs, and production hardening.
Harness-agnostic prompt and agent-configuration API for GitHub Copilot
Supervise Azure DevOps coding agents that run in GitHub Codespaces.
Create concise, Copilot-aware repository instructions, skills, and custom agents.
Adaptive post-task reflection for keeping repository guidance accurate.
[](https://agentpluginsdirectory.com/plugins/project-flood)✓ Verified . We fetched the manifest from GitHub and checked it against the official Agent Plugins 1.0.0 schema at agent-plugins.org.