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 commands/restarter/lets-workflow/startgit clone --depth 1 https://github.com/restarter/lets-workflowWhat 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 | $0.00012 | $0.04170 |
| Opus 5 | $0.00006 | $0.02085 |
| Sonnet 5 | $0.00002 | $0.00834 |
| Haiku 4.5 | $0.00001 | $0.00417 |
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 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.
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 — 298 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Session Start
Restore context and prepare for work. User MUST select a task before working.
LETS Notice — surface it. If a
## LETS Noticeblock is present in the injected context (a one-time message the SessionStart hook emits — e.g. workflow rules outdated/missing, a migration result), it MUST be the first line of your output for/lets:start(⚠️ <notice text>), before the session summary. Don't omit it — it tells the user a/lets:updateor/lets:initis needed. Mention it once, then continue with the steps below; don't repeat it in later turns.
Usage
/lets:start # Full flow - show history, tasks, select work
/lets:start <task-id> # Quick start - jump to specific task
/lets:start --continue # Resume last in_progress task with session context
/lets:start --main # Project-assistant / PM mode - no task, read + triage on merge-branch
/lets:start --assistant # Alias of --main
Step 0: Argument Parsing
If <task-id> provided (e.g., /lets:start lets-rmcwo):
- Skip Steps 1, 3, 5 (session history, orient, task selection)
- Run Step 2 (git state) briefly
- the tracker's
show+comment-listfor<task-id>— read the FULL description and ALL comments, never truncate - Jump to Step 6 (branch) with this task
If --continue:
- Run Step 1 (session history) - important for context recovery
- the tracker's
list-by-status(in_progress) - find task(s) - If exactly 1 in_progress -> use it, skip Step 5
- If multiple -> show selection with context from recent sessions
- If none -> fall through to full flow
If --main or --assistant provided (project-assistant / PM mode):
- Deliberate NO-TASK session stance. Do NOT select, claim, or auto-create a task.
- Precedence: mutually exclusive with
<task-id>and--continue. If an explicit task-id or--continueis ALSO present, the explicit task wins (run the normal task flow) and--mainis ignored - tell the user it was dropped because a task was specified. - Skip Step 5 (Task Selection), Step 6 (Take Task), Step 8 (Task Size Assessment) - all task-bound.
- Run Step 1 (session history), Step 2 (git state). The orient snapshot is rendered once by Main Mode M1 (below), AFTER the session-boundary write - do NOT also run Step 3 (that would render orient twice).
- Then go to
## Main Mode(below) instead of Steps 4-9.
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 · 298 lines · 12 tokens per session scan A 938051a4abbd
start is a command published in the GitHub repository restarter/lets-workflow (17 stars, last pushed 9d ago), licensed MIT. It adds 12 tokens to every session and 4,170 once invoked, about $0.0001 per session on Opus 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 2026-08-30.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.