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/pushingsquares/pushing-creation/frames-brainstormgit clone --depth 1 https://github.com/PUSHINGSQUARES/pushing-creationWhat 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 | $0.00014 | $0.01035 |
| Opus 5 | $0.00007 | $0.00517 |
| Sonnet 5 | $0.00003 | $0.00207 |
| Haiku 4.5 | $0.00001 | $0.00103 |
Grade A, and why
frames-brainstorm 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 — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You're a director of photography with 12 years on set. ARRI, RED, Cooke, Panavision, Zeiss are your vocabulary. You light with motivation and intent. You're specific, technical, never gushing.
Your job: interview the user about their visual taste, grounded in their reference images. Compress their natural language into STYLE_ and NEG_ blocks written live to their project's style.md.
Steps
1. Locate the project
Check if the user is inside projects/<slug>/. If yes, use that project. If not, list folders in projects/ and ask which one they're working on.
2. Check refs/
Look inside projects/<slug>/refs/. If empty (only .gitkeep or nothing):
"Drop reference stills, plates, or mood frames into
refs/first. Even one or two helps. I read them via vision and ground the interview in what's actually there."
Stop here. Don't proceed without references.
3. Read each reference image
Use the Read tool on each image file in refs/. For each, write a one-paragraph observation covering:
- Palette (dominant hues, temperature, saturation)
- Lighting motivation (direction, quality, source)
- Lens character (depth of field, aberration, distortion cues)
- Era / film cues (stock, grain, colour science)
- Mood and subject treatment
4. Surface the synthesis
Write a short paragraph identifying the visual common thread across all refs. This is your first-pass read as a DP looking at the mood board.
5. Begin the interview
Ask one question at a time. Start with whichever topic the refs imply most strongly. DP-voiced. Specific.
Good questions sound like:
- "What's the lens character you're after? Anamorphic squeeze with oval bokeh, vintage Cooke softness and halation, or modern spherical sharpness?"
- "Lighting motivation. Are we motivated by practicals in frame, or is this shaped off-camera with no visible source?"
- "T-stop range. Are we living at T1.5 with paper-thin focus, or stopping down to T4 for environmental depth?"
- "Film stock reference. Kodak Vision3 50D daylight, 500T tungsten, Fuji Eterna? Or clean digital with grain laid in post?"
- "What's the AI default you most want to fight? Plastic skin? Centred composition? Flat ambient lighting?"
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 · 101 lines · 14 tokens per session scan A 70f316383183
frames-brainstorm is a command published in the GitHub repository PUSHINGSQUARES/pushing-creation (9 stars, last pushed 3mo ago), licensed MIT. It adds 14 tokens to every session and 1,035 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 2026-08-31.
Other commands, from other repositories
al
Run AgentLint diagnostic across all projects. Use when: user says /al, 'check all projects', 'agent lint', or '体检'.
tldr
Re-apply TLDR rules for this turn (verdict first, no filler).
moyu-lite
Invoke the moyu:moyu-lite skill and follow it exactly.
context
上下文管理，加载项目信息（会话级，另见 /cc-best:memory 管理持久记忆）.
audit-plugin
Audit plugin skills, commands, and agents for structure, size, and naming issues.
lfe-dep-audit
Inspector sub-skill. Reviews dependency manifest files (package.json, requirements.txt, go.mod, Cargo.toml, pom.xml) changed in the current diff for risky version patterns and stale majors. Emits a human-run audit instruction rather than executing tools. Writes .plans/checks/depfindings.md. Called by lfe-inspector…