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/rmolines/fractal-loop/deliverygit clone --depth 1 https://github.com/rmolines/fractal-loopWhat 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 | $0.00027 | $0.07079 |
| Opus 5 | $0.00014 | $0.03540 |
| Sonnet 5 | $0.00005 | $0.01416 |
| Haiku 4.5 | $0.00003 | $0.00708 |
Grade B, and why
delivery scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
BRANCH_PREFIX=$(grep "^branch-prefix:" .claude/project.md 2>/dev/null | sed 's/^branch-prefix: //' | head -1) How it starts
The opening of the file, as written. The whole thing — 788 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/fractal:delivery
Human gates
Every time this skill needs human input, use the AskUserQuestion tool instead of printing the question as text output.
Context header (REQUIRED on every question when state is available): Prefix the question string with:
📍 | 🎯 <active_predicate (max 80 chars)>
Variables come from the pre-loaded State section. If state is not yet loaded (e.g., early steps of /fractal:propose before tree detection), omit the header.
IMPORTANT: The header must be plain text. No markdown formatting (no **, ##, *, etc.) in the question string. Emojis are fine as visual anchors.
You are the orchestrator (Opus thread) executing an approved plan. Subagents do the implementation. You coordinate, validate, and unblock.
Input: $ARGUMENTS
Core principle
The plan is the contract — execute it faithfully.
Don't improvise deliverables not in the plan. Don't skip validations. Don't re-derive the execution order — read it from the Execution DAG. Subagents never spawn other subagents. All orchestration flows from this thread.
On entry: locate and validate the plan
$ARGUMENTS is the path to the active fractal node directory.
REPO_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
NODE_DIR="${REPO_ROOT}/${ARGUMENTS}"
PLAN="${NODE_DIR}/plan.md"
PREDICATE="${NODE_DIR}/predicate.md"
PRD="${NODE_DIR}/prd.md"
If $ARGUMENTS is empty: read .fractal/root.md → get active_node.
If plan.md not found: "No plan.md found. Run /fractal:planning first."
Read in parallel:
plan.md— especially## Execution DAGpredicate.md— the falsifiable condition for this node (the product reference)prd.md— acceptance criteria (leaf nodes only; provides concrete validation targets).claude/project.md— build/test commands, hot files
Load project context
Scan the .fractal/ tree to understand what has already been built. This provides the agent with awareness of sibling nodes, previous deliveries, and accumulated project state.
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.
- yesterday First seen · 788 lines · 27 tokens per session scan B 11cc1b117870
delivery is a command published in the GitHub repository rmolines/fractal-loop (13 stars, last pushed 5mo ago), licensed MIT. It adds 27 tokens to every session and 7,079 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
pm-arch
Scan project structure and generate Mermaid architecture diagrams + flowcharts under .pm/architecture/.
pm-done
Close a todo (TODO-xxx), sync completed.md, refresh overview.
pm-status
Show governance health, wizard next step, all open blocking/high todos, and pending review count. Primary daily entry.
pm-charter
Create, import, discover, approve, or skip project charter. No-arg form is an interactive wizard.
pm-check
Diagnose .pm/ health (missing files, broken audit, stale map, unconfirmed PRD) and repair or increment without overwriting confirmed content.
pm-tests
List missing unit-test companions for core/changed sources. Do not write tests into the business tree unless the user confirms.