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/thierryn/fire-flow/fire-execute-plangit clone --depth 1 https://github.com/ThierryN/fire-flowWhat 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 | $0.00017 | $0.01282 |
| Opus 5 | $0.00009 | $0.00641 |
| Sonnet 5 | $0.00003 | $0.00256 |
| Haiku 4.5 | $0.00002 | $0.00128 |
Grade A, and why
fire-execute-plan scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 183 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/fire-execute-plan
Execute one BLUEPRINT.md with intelligent segmentation, atomic commits per task, and enforced test gates.
Arguments
arguments:
plan_id:
required: false
type: string
description: "Plan ID (e.g., 03-02). If omitted, auto-detects next unexecuted plan."
optional_flags:
--continue: "Resume from .continue-here.md"
--branch: "Custom feature branch name"
--no-branch: "Execute on current branch"
--dry-run: "Parse segments without executing"
--main-only: "Force all segments to main context"
Process
Step 1: Load Project State
+---------------------------------------------------------------+
| DOMINION FLOW >>> PLAN EXECUTION |
+---------------------------------------------------------------+
Load CONSCIENCE.md: current phase, accumulated decisions, blockers/concerns.
Step 2: Identify Plan
If plan_id provided: locate directly. If omitted: find first BLUEPRINT.md without matching RECORD.md.
Step 3: Create Feature Branch
Unless --no-branch:
git checkout -b "feat/${PHASE}-${PLAN}-${PLAN_NAME_SLUG}"
Step 4: Parse Segments and Select Execution Mode
Parse segments based on checkpoints:
| Pattern | Condition | Segment Strategy |
|---|---|---|
| A - Autonomous | No checkpoints | Single executor, entire plan |
| B - Segmented | Verify-only checkpoints | Executor per segment, main for checkpoints |
| C - Decision-Dependent | Decision/action checkpoints | All in main context |
Auto-select execution mode (see @references/execution-mode-intelligence.md):
auto_tasks = tasks where type == "auto"
checkpoint_tasks = tasks where type starts with "checkpoint:"
specialties = detect_specialties(auto_tasks)
IF checkpoint_tasks has "checkpoint:decision":
MODE = SEQUENTIAL
REASON: "Decision checkpoint requires user interaction"
ELIF auto_tasks.count >= 3 AND tasks_are_independent(auto_tasks)
AND plan.risk_level != "high":
MODE = SWARM
REASON: "3+ independent tasks — Team Lead delegates to specialists"
TEAM: compose from detected specialties (backend/frontend/test)
ELIF auto_tasks.count >= 2 AND tasks_are_independent(auto_tasks):
MODE = SUBAGENT
REASON: "2 independent segments — parallel subagent execution"
ELIF plan.risk_level == "high":
MODE = SEQUENTIAL
REASON: "High-risk plan — careful serial execution"
ELSE:
MODE = SEQUENTIAL (default)
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 · 183 lines · 17 tokens per session scan A 80f4c43f8294
fire-execute-plan is a command published in the GitHub repository ThierryN/fire-flow (77 stars, last pushed 20d ago), licensed MIT. It adds 17 tokens to every session and 1,282 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
template
Manage issue templates for streamlined issue creation.
project-recap
Generate a visual project recap for context switching.
speckit.taskstoissues
Convert existing tasks into actionable, dependency-ordered GitHub issues for the feature based on available design artifacts.
ado-pull
Pull latest changes from Azure DevOps (like git pull). Supports increment, project, or full living docs sync.
issue-amend
Re-snapshot the active issue's scope from the spec, clear verified and reviewed receipts, and record the change as a permanent amendment.
dispatch-worker
Run one in-session-agent Worker tick (RFC-0041 §4.3.1).