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/simonrowland/goal-flight/validate-queuegit clone --depth 1 https://github.com/simonrowland/goal-flightWhat 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 | $0.00007 | $0.01246 |
| Opus 5 | $0.00003 | $0.00623 |
| Sonnet 5 | $0.00001 | $0.00249 |
| Haiku 4.5 | $0.00001 | $0.00125 |
Grade A, and why
validate-queue 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 — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
validate-queue []
Schema-check a goal-queue file. Read-only. Surfaces structural problems before they bite mid-execute.
When to invoke
- The user explicitly typed
/goal-flight validate-queue [<path>]. - After hand-editing a goal-queue and before running
/goal-flight execute. - Pre-flight inside
commands/execute.mdstep 1 (called automatically; bail with a clear error if validation fails).
What the user provides
- No args → validate the most recent
docs-private/goal-queue-*.md(new naming as of 0.3.0); fall back to legacydocs-private/<topic>-goal-queue-*.mdif no new-form file exists. - One arg
<queue-file>→ validate that specific path.
Steps
-
Resolve the queue file (most recent by date suffix if no arg).
-
Parse the queue. Each chunk is a
## #N <slug>heading followed by four mandatory subsections (per the goal-queue shape incommands/decompose-plan.md):### SCOPE— one paragraph.### CHECKLIST— bulleted list.### ACCEPTANCE— testable criteria, bulleted.### FORBIDDEN— explicit anti-scope, bulleted.
-
Run these checks. Each failure is a P0/P1/P2:
Structural (P0 — blocks execute):
- Every chunk heading matches
^## #(\d+) (.+)$— extract chunk number + slug. - Numbering is sequential starting from 1. No gaps. No duplicates.
- Each chunk has all four mandatory subsections (
SCOPE,CHECKLIST,ACCEPTANCE,FORBIDDEN). None are empty. - Slugs are unique across the queue.
- Slugs are lowercase + hyphens only (no spaces, no underscores, no caps).
Tags (P1 — execute can proceed but flag):
[parallel-safe:<group>]tags reference groups that have ≥ 2 members. A lone[parallel-safe:foo]chunk is parallelism-broken and should be untagged or grouped.[milestone]tags should appear roughly every 5 chunks for projects ≥ 15 chunks; warn if absent.[controller-direct]is a candidate tag, not permission. SCOPE must show held session context and a fully stateable edit with no exploration or special review-independence need. Warn on unread call sites, cross-module discovery, or a non-mechanical trust surface — not delta size or file count. Recheck free-slot serialization under Axis 2 at execution.[acp]and[bash-tail]are mutually exclusive — co-occurrence on the same chunk → P1 conflict warn (orchestrator would otherwise have to pick arbitrarily).[acp]should only appear whendocs-private/env-caveats.mdshows the target adapter installed (otherwise the dispatch will fall through to[bash-tail]at runtime; warn if env-caveats is missing OR shows the adapter unavailable — execute can still proceed via fallback).[goal-mode] + [bash-tail]co-occurrence is codex-only — only codex/goalemits the Final-response marker the watcher needs to detect end-of-goal in a flat tail. This rule is P0 (escalated from the surrounding P1 block): if a chunk has[goal-mode] + [bash-tail]AND the target worker is not codex (or worker isn't specified and would default to non-codex per env), the dispatch will hang waiting for a marker that never comes. Seeprotocols/dispatch-routing.mdcomposition table for the rule.
- Every chunk heading matches
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 · 106 lines · 7 tokens per session scan A eb9ef8b15871
validate-queue is a command published in the GitHub repository simonrowland/goal-flight (20 stars, last pushed 4d ago), licensed MIT. It adds 7 tokens to every session and 1,246 once invoked, about $0.0000 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
review-context7-updates
Please read the @README.md and @docs/tools/package-documentation.md to understand the high level context of my repo.
init
Set up Memento — configure MCP server and verify everything works.
mine
Quickly capture relevant project context into Memento memory.
help
Show comprehensive Memento help — available MCP tools and usage patterns.
search
Search your memories using Memento hybrid retrieval.
status
Show Memento server status and current configuration.