By tyroneross
Agent Rally Point — repo-local coordination surface for parallel coding agents (Claude Code, Codex, Gemini). Records claims, checks write boundaries, routes handoffs, and exposes room state via the rally CLI. Skills: agent-rally-point (per-session loop), rally-workflows (fan-out + dynamic-workflows protocol), mini-loop (per-task quality gate). Advisory-only hooks auto-fire on SessionStart + PreToolUse to surface room awareness and deconflict before edits — never blocks. Backend: rally CLI (cargo install --path crates/rally-cli).
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Use when working in a repository that uses Rally/Agent Rally Point for cross-agent coordination, especially at session start, before editing files, when deciding what to do next, handing work to another agent, recording facts/artifacts/decisions, resolving blockers, or coordinating with other coding agents through the `rally` CLI.
Ultra-light per-task quality loop for Rally Flow / dynamic-workflows fan-out. Wrap each fanned-out task in assess → plan → execute → mini-judge to raise accuracy, with zero tools or scripts. Use when running a workstream task and you want it to actually meet its validation + output contract, not just "be done".
Use when fanning out work across multiple agents, running a dynamic workflow, coordinating parallel subagents, or splitting a workstream across hosts, terminals, or machines through Agent Rally Point. Defines the workstream descriptor + task-packet protocol and the per-task rally coordination loop. Host-neutral — works for any coding agent.
Modifies files
Hook triggers on file write and edit operations
Rally is a repo-local coordination room for coding agents working in the same checkout: durable facts, current room state, next-action guidance, and write-boundary checks without a server.
The only shipped coordination command is rally.
claude plugin marketplace add tyroneross/agent-rally-point
claude plugin install agent-rally-point@agent-rally-point
Hooks (SessionStart auto-coordination + PreToolUse write-boundary checks) and skills (agent-rally-point, rally-workflows, mini-loop) activate automatically on install. The rally CLI binary is auto-provisioned on first session.
New agent or human dropping in? Read RALLY.md. It is the short
operating guide for the Rally loop.
The full product boundary is docs/RALLY_ARCHITECTURE.md.
The long-term coordination protocol model is
docs/PROTOCOL-NORTH-STAR.md.
Command read/write behavior is summarized in
docs/COMMAND-SEMANTICS.md.
The build-loop implementation and Claude/Codex dogfood plan is
docs/PLAN-protocol-claim-authority-dogfood.md.
Checkout and active-ledger migration rules live in
docs/CANONICAL-CHECKOUT-MIGRATION.md.
Automatic Claude Code / Codex hook wiring (SessionStart + PreToolUse) is
documented in docs/AUTO-COORDINATION-HOOKS.md.
Rally owns the primary product path:
rally is the primary CLI..rally/log/<engagement>.jsonl is the durable fact store, including managed
sessions..rally/facts.db is a derived sqlite cache rebuilt from the log.enter, next, say, room, check, run, sessions, inject,
attach, capture, stop, locate, recent, lead, backlog, and
mission are the load-bearing commands.Network transport remains out of scope. Files, Git, rsync, shared folders, A2A, or a future service can move facts; Rally defines what the bytes mean.
Install the primary Rally CLI from the checkout:
git clone https://github.com/tyroneross/agent-rally-point.git
cd agent-rally-point
cargo install --path crates/rally-cli
rally whoami --tool codex --json
rally enter --tool codex --json
rally ack --tool codex
rally next --tool codex --json
The primary loop is intentionally small:
rally whoami --tool codex --json
rally enter --tool codex --json
rally ack --tool codex
rally next --tool codex --json
rally check before-write --tool codex --path crates/foo.rs --strict --json
rally say artifact --tool codex --subject "implementation complete" --uri crates/foo.rs --evidence "cargo test" --json
rally room --json
The autonomous act-on-next contract is:
whoami -> enter -> ack -> next -> if actionable, claim/check -> execute
-> verify -> say artifact/handoff/resolve/release -> next
rally next returns actionable, requires_human, stop_reason,
suggested_claims, suggested_commands, and completion so an agent harness
can act without turning Rally into a scheduler or coding agent.
Useful fact writes:
rally say claim --tool codex --subject "edit parser" --path crates/rally-cli/src/main.rs --json
rally say release --tool codex --ref <claim-id> --subject "done" --json
rally say blocker --tool codex --subject "need merge decision" --severity high --json
rally say resolve --tool codex --ref <blocker-id> --subject "resolved" --json
rally say decision --tool codex --subject "Rally is primary" --status binding --json
rally say handoff --tool codex --target claude_code --subject "review docs" --summary "Rally is now primary" --json
Resolve handoff targets from live Rally state, not from examples or old logs.
Use rally whoami, rally lead show, rally next, rally room --json, and
explicit handoff targets to identify the current recipient. A targeted handoff
is the durable action request; rally inject is only a wake/delivery path for a
target already listed by rally sessions --json.
For --handoff, rally inject waits for target-authored Rally evidence by
default. Text reaching a pane is not proof that the receiving agent read or
acted on it. If the target does not post resolve, receipt, artifact,
blocker, or decision for the expected handoff before the timeout, the
inject result returns ack_state: "timeout", verified_received: false, and a
fallback_plan. Treat that as not received: check Rally state and assigned
file movement, retry once with a short doorbell, then move the work to a
separate worktree, hand it to another live agent, or escalate.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimnpx claudepluginhub tyroneross/agent-rally-point --plugin agent-rally-pointBarbara Minto's Pyramid Principle for short-form, long-form, presentations, and audits
Coding Debugger for coding agents with debugging memory, verdict-based retrieval, pattern extraction, root-cause workflows, and parallel domain assessment.
Map dependencies and analyze blast radius before changing code
End-to-end design tool for AI coding agents — guided UI builds, web/iOS/macOS guidance, visual checks, mockup matching, interaction testing, native scanning, native layout-fill / gap analysis
Multi-phase build orchestration with native debugging: assess, plan, execute, review, iterate. Plan-verify gate on Phase 2. Optional UI design gates. Native debug-loop, debugging-memory, and logging-tracer skills provide root-cause investigation without a bundled MCP server. Orchestrator owns when-to-fire; bundled skills own procedural detail; logging-tracer-bridge is an optional escalation hop to the standalone Coding Debugger supporting plugin for extended capability.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.