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.
git clone --depth 1 https://github.com/frankxai/Starlight-Intelligence-SystemWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/frankxai/starlight-intelligence-system/align-voice)<a href="https://agentmods.dev/commands/frankxai/starlight-intelligence-system/align-voice"><img src="https://agentmods.dev/badge/commands/frankxai/starlight-intelligence-system/align-voice.svg" alt="Measured on agentmods" height="20"></a>What 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 5.1 | $0.00052 | $0.01893 |
| Opus 5 | $0.00026 | $0.00946 |
| Sonnet 5 | $0.00010 | $0.00379 |
| Haiku 4.5 | $0.00005 | $0.00189 |
Grade A, and why
align-voice 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 7d 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 — 177 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/align-voice
Load SIP.md, VOICES.md, agents/starlight-visionary.md, skills/vision/design-coherence.md. Audit existing content against the person's Brand Kit + Genius Profile voice samples. Produce a Voice Audit document with score, drift patterns, rewrite examples, and guardrails.
Input
$ARGUMENTS
When this command fires
- Brand Kit exists at
vision/brand-<slug>.md - Content exists that pre-dates the Brand Kit or was drafted by a third party (agency, ghostwriter, AI without Profile)
/creator-pipelineis about to ship — voice-align is the gate- The person says "this doesn't sound like me" about their own published work
When this command does NOT fire
- No Brand Kit → halt and route to
/build-brand-kit(no target to align to) - No content submitted → halt and request content
- Single-piece alignment that only needs a quick rewrite → use
brand-voiceskill directly
Process
-
Load upstream.
- Resolve
<person-slug>. - Load
genius/profile-<slug>.md(voice samples + distinctive vocabulary). - Load
vision/brand-<slug>.md(voice rules + visual mood + don'ts). - If any missing, halt and route to upstream command.
- Resolve
-
Collect content.
- Accept: file paths, pasted text blocks, URLs to public posts, email drafts, landing copy.
- For each item, record: source, date (if known), type (essay / social post / email / landing / other).
- Require ≥ 3 items for
--depth full, ≥ 1 item for--depth quick.
-
Score each item against voice rules.
- For each voice rule from Brand Kit, check content: pass / drift / fail.
- Compute coherence score per item = passes / (passes + drifts + fails) expressed as 0.00–1.00.
- Compute aggregate score = average across items (with fails weighted 2x).
-
Identify drift patterns.
- Look for recurring drift across items. Examples:
- "Reverts to corporate-we in transitions" — drift from rule "first-person only"
- "Uses 'leverage' and 'synergies'" — drift from distinctive vocabulary (Profile never uses these)
- "Listicle structure in long-form" — drift from rule "one framework per piece"
- Name each pattern specifically. Cite the items where it appears.
- Look for recurring drift across items. Examples:
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.
- 7d ago First seen · 177 lines · 52 tokens per session scan A 6e7ff1a202d2
align-voice is a command published in the GitHub repository frankxai/Starlight-Intelligence-System (8 stars, last pushed yesterday), licensed MIT. It adds 52 tokens to every session and 1,893 once invoked, about $0.0003 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 2026-08-31.
Other commands, from other repositories
memories
View and manage learned memories.
mpm-session-resume
Load context from paused session.
forget
Delete specific memories.
learn
Add new learning to memory.
gbu-retro
Post-session retrospective — harvest this session's lessons into durable doctrine.
consolidate
Write a compact checkpoint summary of the current frontier.