Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/cuecrux/crux/execplan-runnpx skills add CueCrux/Crux --skill execplan-rungit clone --depth 1 https://github.com/CueCrux/CruxWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00174 | $0.01692 |
| Opus 5 | $0.00087 | $0.00846 |
| Sonnet 5 | $0.00035 | $0.00338 |
| Haiku 4.5 | $0.00017 | $0.00169 |
Grade A, and why
execplan-run scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
compact block. Use it instead of hand-rolling `curl | jq` or `gh` loops — that is the whole How it starts
The opening of the file, as written. The whole thing — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ExecPlan Run
The four execplan-* skills author plans. None of them close one. That is why 63 plans
sit in_progress and 25 of those have zero milestones gated (measured 2026-08-06). This
skill is the execution and closeout loop.
Announce at start: "Using execplan-run — ." Then work the phases.
Two lanes
| Lane | Who | Reads |
|---|---|---|
| Orchestrator — pick, check collisions, dispatch, shepherd, close | the session working with the operator | references/orchestrator.md |
| Worker — one milestone, gate to gate | a worktree session or subagent | references/milestone-loop.md |
Default to orchestrator unless handed a specific <slug> M<n> to implement.
Closeout is the orchestrator's job and it is not optional:
references/closeout.md.
Mechanics live in a script, not in your context
scripts/ep does every board read, coordination call, worktree op, and PR poll and prints a
compact block. Use it instead of hand-rolling curl | jq or gh loops — that is the whole
token argument.
ep board [n] # ranked open work, already in recommended order (~650 tok)
ep preflight <slug> # grounding + board row + live peers + worktrees, one block
ep announce <slug> [M<n>] [paths] # declare focus; prints overlaps. ttl 0 clears
ep worktree <repo> <slug> [branch] # worktree off origin/main, named for the slug
ep pr open|watch|merge <dir> [...] # push, poll checks, squash-merge
ep reap [--dry] [repo] # delete worktrees whose branch already merged
Facts stay on MCP (store_fact / query_facts) — the script never writes them, because
facts must be deliberate.
The loop
0 BOOT once per session — banner, or ep board
1 PICK ep board → take the FIRST entry (it is already ranked)
2 PREFLIGHT ep preflight <slug> → grounding, blockers, peers, ODs
3 CLAIM ep announce + ep worktree + save_session
4 MILESTONE implement → test → commit → store_fact gate:M<n> ← repeats
5 PR ep pr open → watch → merge (automatic; no menu unless the plan changed)
6 CLOSE last milestone → Status flip, OD sweep, ep reap, announce ttl=0
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 128 lines · 174 tokens per session scan A 5e9e050f118e
execplan-run is a skill published in the GitHub repository CueCrux/Crux (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 174 tokens to every session and 1,692 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
capture-task
Capture a new task — create a draft or pending Task record from a rough idea or detailed spec. Use when asked to "add a task", "log a bug", "create a task", or "add to backlog".
refine-backlog
Refine the backlog — screen drafts, remove duplicates, fill in missing details, classify, link, and move to pending.
forward-plan
Use when turning Brigade's native ready set, outcome rank, and ROADMAP.md into a dependency-filed plan artifact under .brigade/work/plans/ - propose GraphTrail-derived edges (blocks from def/use, conflicts-with from write overlap) marked derived and requiring confirmation. Degrade cleanly when GraphTrail is…
implement-task
Pick up and work on a binder Task — investigate bugs, discuss open questions, design solutions, write tests, or implement code. Use when asked to "work on", "pick up", "implement", or "fix" a task.
handoff-inbox-drain
Use when draining a Brigade handoff review inbox - triage drafts with allowlist-bound Brigade handoff commands, grounded edits that cite on-disk content, archive-only retention, and reinforce-in-place when ingest already matched a known fact (#724). Never delete, remove, or prune.
phase-orchestrator
Orchestrates phase-based implementation plans using the host's todo tracker and subagents for implementation, review, validation, default phase commits, plan finalization, and a final end-to-end review. Delegates work through the companion phase-implementer, phase-reviewer, and phase-final-reviewer skills and consumes…