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/dsifry/metaswarm/start-taskgit clone --depth 1 https://github.com/dsifry/metaswarmWhat 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 | $0.00000 | $0.01852 |
| Opus 5 | $0.00000 | $0.00926 |
| Sonnet 5 | $0.00000 | $0.00370 |
| Haiku 4.5 | $0.00000 | $0.00185 |
Grade A, and why
start-task 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 — 197 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Start Task
Determine task complexity and use appropriate workflow for efficient development.
Usage
/start-task <task-description>
Steps
0. Pre-Task Checklist
Before starting any new task:
Context Recovery Check:
- Check if
.beads/plans/active-plan.mdexists withstatus: in-progress - If YES: An interrupted execution exists. Ask user: "There's an active plan from a previous session. Resume it or start fresh?"
- Resume → run
bd prime --work-type recoveryand pick up where execution stopped - Start fresh → mark the old plan as
status: abandonedand proceed normally
- Resume → run
Knowledge Priming (CRITICAL):
- Run BEADS prime:
bd prime --keywords "<task-keywords>" --work-type planning - Review MUST FOLLOW rules and GOTCHAS before proceeding
- Note any relevant patterns or decisions that constrain the approach
PR Check:
- Check if there are active PRs with pending comments
- Ask: "Before we start the new task, should we check if there are any PR comments to address?"
- If yes, run:
gh pr list --author @me --state open - Check each PR for new CodeRabbit comments
0.5. External Tools Check (informational only — never blocks the task)
Check if external AI tools (Codex, Gemini) are available for cost savings and cross-model review:
- Run the health check scripts to detect installed tools:
command -v codex >/dev/null 2>&1 && echo "codex: available" || echo "codex: not found" command -v gemini >/dev/null 2>&1 && echo "gemini: available" || echo "gemini: not found" - If tools are detected but
.metaswarm/external-tools.yamldoes not exist: Suggest the user enable them:"External tools (Codex/Gemini) are installed but not configured. Run
mkdir -p .metaswarm && cp templates/external-tools.yaml .metaswarm/to enable cost-saving delegation." - If no tools are detected: Briefly mention they can be installed:
"Optional: Install Codex and Gemini CLIs for cost savings and cross-model review — see
templates/external-tools-setup.md." - If tools are configured and working: No message needed — proceed silently.
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 · 197 lines · 0 tokens per session scan A 4478b3143c73
start-task is a command published in the GitHub repository dsifry/metaswarm (407 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,852 tokens. 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
git
Git operations with intelligent commit messages and workflow optimization.
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.