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 skills add Tradebaas/Groundwork --skill startgit clone --depth 1 https://github.com/Tradebaas/GroundworkWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tradebaas/groundwork/start)<a href="https://agentmods.dev/skills/tradebaas/groundwork/start"><img src="https://agentmods.dev/badge/skills/tradebaas/groundwork/start/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tradebaas/groundwork/start"><img src="https://agentmods.dev/badge/skills/tradebaas/groundwork/start.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00081 | $0.00732 |
| Opus 5.5 | $0.00032 | $0.00293 |
| Sonnet 5.5 | $0.00016 | $0.00146 |
| Haiku 4.5 | $0.00008 | $0.00073 |
Grade A, and why
start scanned grade A with 0 findings 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 5d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 43 lines — stays where its author put it; the contents beside it link to each section on GitHub.
start: one ticket, one session, the right shape
Model and effort are set once, globally, and never switched mid-session: each model has its own prompt cache, and a switch re-reads the whole history at full price. What this skill picks is the shape of the session around that fixed model.
1. Where we are
docs/state/STATE.mdexists: read its handoff block and nothing else; it is the starting point, checked against the latest commits before you act onNow ▶.- No STATE.md: ask in one line what this session is for. Do not propose a governance layer;
beginlays one down only when the owner asks for it.
2. The ticket
One ticket per session. If the owner names more than one, take the first and park the rest in
STATE.md's Now ▶ chain or in INTAKE.md. When the ticket is done: handoff, then /clear.
3. The shape, from this table
| Situation | Do this | Not this |
|---|---|---|
| The diff cannot be described in one sentence | plan mode first (Shift+Tab); the owner approves the plan | build and roll back |
| "Keep going until the tests pass" or any condition the output can prove | /goal "<condition>, or stop after 20 turns", in auto mode |
a /loop that resends the whole context every few minutes |
| Waiting on CI, a deploy, a remote job | /loop 8m check <what>, or the Monitor tool when the session has it |
polling by hand, or a loop under the cache lifetime |
| Recurring work that must run without this machine | a routine via /schedule, on Sonnet, with a self-contained prompt |
a local /loop, which expires after 7 days |
| A hard call where plan quality decides the outcome | /advisor opus beside the session model; toggling it keeps the cache |
opusplan or /model mid-session |
| Reading logs, test output, many files | a subagent on Sonnet that returns the conclusion | pulling the output into this context |
| Two failed corrections on the same problem | /clear with a better first prompt, or /rewind to the turn before the dead end |
a third attempt in a context full of the first two |
| Fable | only for the hard tier, chosen before the session starts | as the default |
Agent teams, ultracode |
only for work that is genuinely parallel and independent | as a speed-up for ordinary tickets |
A plugin skill that pins its own model |
start the session on that model (Code Apps skills pin Sonnet, Power Automate skills pin Opus) | invoking it from a session on another model: every call is a model switch |
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.
- 5d ago First seen · 43 lines · 0 tokens per session scan A b732954971c8
start is a skill published in the GitHub repository Tradebaas/Groundwork (2 stars, last pushed 6d ago), licensed MIT. It adds 81 tokens to every session and 732 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-11.
Other skills, from other repositories
keel-handover
Add this session's block to HANDOVER.md before ending or compacting — done / tried-failed / latest / next. Rotates via /keel-distill when the 3-block cap is hit.
consolidate-interview
End of session — replace running notes with the full synthesis across all three funnel levels and author the glossary of ubiquitous language.
publish-release
Publish the delivered batch as release notes, a PR to main, or a tagged release.
agentmail
Use your assigned AgentMail inbox to read email tasks, explicitly send or reply, and check delivery. Provided automatically by your inbox assignment.
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
Cortex
Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…