donaldfilimon/abi / .agents/skills/goals
goals
VerifiedCapture, track, and execute goals for the abi stable-Rust project. Use when the user wants to record a goal, see outstanding goals, update goal status, or break a goal into a verified work slice. Distinct from abi-goal-orchestrator (which plans from the TODO board) — this skill owns a persistent goal ledger and drives goal-to-slice execution.
Goals
A lightweight goal ledger for the abi project plus a loop that turns a goal into
a verifiable work slice and records the outcome. Goals persist in
tasks/goals.md so they survive sessions; the active board (tasks/todo.md)
remains the granular working list.
Ledger
The ledger is tasks/goals.md, a markdown file with one ## <Goal> section per
goal. Each section carries a status line and a short body:
## Ship file-aware agent context
status: done
- Wired @file resolution into agent plan/multi + TUI repl
Valid status: values: todo, in_progress, blocked, done.
If tasks/goals.md does not exist, create it with a # Goals header. Treat the
file as the source of truth; never delete a closed goal — move a one-line note
to the bottom or mark done.
Workflow
- Start in the abi repo root unless the user gives another checkout.
- Inspect
git status --short --branchbefore edits; preserve unrelated dirty work. - Read
AGENTS.md,tasks/todo.md, andtasks/lessons.mdfor conventions. - Capture: if the user states a goal, ensure it has a
## <Goal>section intasks/goals.mdwithstatus: todo(orin_progressif started this turn). - Track: list goals by reading
tasks/goals.md.update <goal> <status>rewrites that goal'sstatus:line. - Execute: derive the smallest slice that moves the goal forward. Prefer a change that makes one TODO, roadmap gap, doc mismatch, or validation gap measurably more true. Implement it, then verify.
- Keep claims honest: source/build/tests override prose. Update the goal's status and append a one-line outcome note when the slice lands.
Goal Rules
- Do not convert disclosed stubs into fake completions. Native GPU kernels, production clustering, production FHE, and learned-compression claims need real source/tests.
- Do not add legacy CLI names. Preserve the frozen top-level command set and the MCP 12-tool contract.
- When changing docs, keep claims aligned with
docs/contracts/external-claims-audit.mdx. - Goals are user intentions, not tasks — keep
tasks/goals.mdcoarse and lettasks/todo.mdhold the granular steps.
Verification Ladder
Prove the slice with the narrow command first, then broaden:
./tools/cargo.sh test -p <crate> --lib -- <filter>
./tools/check.sh
Smoke the real binaries after surface changes:
./tools/cargo.sh build -p abi-cli -p abi-mcp
./target/debug/abi backends
./target/debug/abi complete "hi"
printf '%s\n' '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' | ./target/debug/abi-mcp 2>/dev/null
Always use ./tools/cargo.sh — never bare Homebrew cargo.
Notes
- This skill is project-scoped (lives under
.agents/skills/goals/); it does not change the CLI surface or MCP tools. - Mirrors the taxonomy and honesty constraints of
abi-goal-orchestratorbut owns the persistent ledger rather than reading the TODO board only.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |