SKILL: Lofn Director — Video Prompt Writer

⛔ YOU DO NOT RENDER VIDEOS

lofn-director writes video prompts (for Veo, Kling, Wan, etc.). It does NOT call any video generation API or tool.

Do NOT use: video_generate, veo, or any video generation tool. Do NOT attempt to produce video files.

Your output is: .md files containing final ranked video prompts ready for rendering. The main session or cron agent handles actual Veo/Kling API calls.

---

🔴 AGENT SAVE-OUT PROTOCOL — MANDATORY

Every subagent must save artifacts to disk at EACH major pipeline step, not just at the end. If an agent times out or the session resets, partial work must already be on disk.

Write step files as you complete them. Never hold all output in memory for a single final write.

---

🔗 CREATIVE CONTEXT — FULL CONTEXT INJECTION (MANDATORY)

Every step (00–10) MUST receive the COMPLETE upstream context — no compression, no summarization. The orchestrator fills OVERALL_PROMPT_TEMPLATE.md once from its handoff artifacts (02_golden_seed.md, 03_orchestrator_panel_debate.md, 04_orchestrator_metaprompt.md, 06_*_handoff.md) into a single CREATIVE CONTEXT block: user request + Golden Seed + meta-prompt + personality + all 3 panels (18 voices) + 15 Special Flairs. That block is injected verbatim into the CREATIVE CONTEXT slot at the top of EVERY step file under steps/, alongside each step's prior outputs. Use the supplied Panel Ledger — do NOT invent a new panel; each Devil's Advocate / Hyper-Skeptic must dissent.

---

🔴 SPLIT-STEP AGENT ARCHITECTURE

Do not use the legacy lofn-director agent for Steps 06-10. Use dedicated step agents:

Step	Agent ID	Model
00-05	lofn-director-coordinator	deepseek/deepseek-v4-pro
06	lofn-director-step06	openrouter/qwen/qwen3.7-max
07	lofn-director-step07	openrouter/qwen/qwen3.7-max
08	lofn-director-step08	openrouter/qwen/qwen3.7-max
09	lofn-director-step09	openrouter/qwen/qwen3.7-max
10	lofn-director-step10	deepseek/deepseek-v4-pro
11	lofn-director-step11	openai/gpt-5.5

Full spec: vault/DIRECTOR_MODEL_ASSIGNMENTS.md

---

🧬 PERSONALITY INJECTION MANDATE

Every creative pipeline task MUST receive the target personality's full DNA block — not just "use the Solarpunk Bloom aesthetic."

Saying "style = X" or "use Industrial Grief mode" is insufficient — subagents default to Lofn's system-context personality. The personality block MUST include: cinematic identity, camera language preference, light/shadow philosophy, motion signature, and core directorial beliefs.

Inject this block at EVERY stage: seed, orchestrator, director coordinator, pair agents, step11. Without injection: Lofn bleed. With injection: target cinematic identity.

---

🔴 PIPELINE POSITION: PHASE 2 — YOU ARE NOT FIRST

The correct pipeline order is: Research → Lofn-Core → Orchestrator → YOU → QA

You should be receiving an orchestrator-metaprompt.md and orchestrator-brief.md before you run. If you do not have these files in your output directory:

• Check if core-brief.md exists (Lofn-Core ran but orchestrator skipped)
• If neither exists, you are being invoked out of order - flag it and proceed with best-effort
• If both exist, you are correctly positioned - proceed with the full pipeline

Lofn-Core's job: seed + research. Orchestrator's job: panel + personality + metaprompt. Your job: steps 00-10, per-pair execution, renders.

---

🔴 PRE-CREATIVE ORCHESTRATOR PACKET GATE

Before this modality agent begins Step 00, it must validate a real Lofn-Core + orchestrator packet with:

python3 /data/.openclaw/workspace/scripts/validate_orchestrator_packet.py <run_dir>

The packet must use the original Lofn panel-object structure: Special Flairs, Concept Panel, Medium Panel, and Context & Marketing Panel, each with a Devil's Advocate / Hyper-Skeptic adversarial role. If validation fails, do not proceed; request/launch lofn-orchestrator work.

Every canonical step artifact must use /data/.openclaw/workspace/scripts/lofn_step_artifact_template.md and pass validate_with_retries.py before the next step.

---

⚡ MANDATORY SUBAGENT SPLIT ARCHITECTURE

YOU MUST ALWAYS USE THIS PATTERN. Never run all 10 steps in a single agent.

See TASK_TEMPLATE.md for the full specification. Summary:

When spawned as the "director coordinator":

1. Run steps 00-05 ONLY → produce concept_medium_pairs.json
2. Report back to main session with the 6 pairs
3. Main session spawns 6 parallel pair subagents (steps 06-10, one per pair)

When spawned as a "pair agent" (steps 06-10):

• You will receive ONE concept-medium pair via the lean pair-agent input standard in /data/.openclaw/workspace/vault/LEAN_PAIR_AGENT_INPUT_STANDARD.md
• Run steps 06-10 for that pair only
• Output 4 shot prompt variants to step10_final_pair{N}.md
• Return as completion message

Spawn Pattern (one step at a time, proven in music):

After Step 05 completes:

1. Spawn up to 5 pair agents for Step 06 only using lofn-director-step06
2. Verify files on disk. Validate. Then spawn Step 07 agents using lofn-director-step07
3. Continue one step at a time through Step 10
4. Step 11 enhancement uses lofn-director-step11

Why one step at a time: Running Step 06-10 in a single agent per pair was the #1 cause of timeout churn and template collapse in music.

Why: A single agent cannot faithfully execute all 10 steps without collapsing into templates. This was proven across 3 failed runs (2026-03-30). The split is the fix. It matches the original Lofn ui.py architecture exactly.

Stricter correction (2026-05-20): The split alone is not enough. Each numbered step must be a separate model call with a separate canonical artifact. Coordinator summary files and per-pair omnibus files are pipeline violations.

Lean pair-input correction (2026-05-21): Pair agents must not be handed the whole upstream packet by default. The parent/controller validates the full packet, then gives each video/animation pair agent a compact 50–100 line brief: compact Golden Seed operating excerpt, Step 05 artifact, concept_medium_pairs.json, one pair assignment excerpt, video Step 06–10 contract, tiny provenance block, and motion/model/safety blockers. This is mandatory for video and animation pair work.

---

PREREQUISITES: 0. Load resources/panel-of-experts.md to understand the panel of experts prompting you will use.

1. Load skills/lofn-core/SKILL.md for personality and Panel system.
2. Load skills/lofn-core/PIPELINE.md for the MANDATORY execution pipeline.
3. Load skills/lofn-core/OUTPUT.md for the MANDATORY artifact saving format.
4. Load skills/video/TASK_TEMPLATE.md for exact output requirements. 4a. For pair-agent spawning or pair-agent execution, load /data/.openclaw/workspace/vault/LEAN_PAIR_AGENT_INPUT_STANDARD.md and use the compact pair brief; do not load the full upstream packet into ordinary pair agents.

⚠️ EVERY video generation MUST follow the full pipeline: 10 steps, 3 panels, 6 pairs × 4 outputs = 24 treatments minimum. Then select the best N to return. No shortcuts.

---

🎬 PURPOSE

Transform video concepts into production-ready prompts for video generation (Veo 3.1, Runway, Kling, Pika). This skill handles cinematic direction from concept to viral-ready output.

---

🎲 RANDOM INJECTION (MANDATORY PRE-STEP)

Before executing Step 00, run the random injection script to populate creative seeds:

node /root/.openclaw/workspace/scripts/random-injection.cjs video 50

This returns JSON with:

• aesthetics - 50 random items from the master list (3000+ aesthetics)
• frames - 50 random items from video_frames.csv (1272 cinematic framing techniques)
• film_styles - 50 random film styles for cinematic direction

Inject these into Step 00's placeholders to "shake up the creative space."

---

📊 PIPELINE OVERVIEW

Step	File	Output
00	00_Generate_Video_Aesthetics_And_Genres.md	50 aesthetics, 50 emotions, 50 frames, 50 genres
01	01_Generate_Video_Essence_And_Facets.md	Essence + cinematic style axes + 5 facets
02	02_Generate_Video_Concepts.md	12 distinct concepts
03	03_Generate_Video_Artist_And_Critique.md	Director influence + critique per concept
04	04_Generate_Video_Medium.md	Visual style assignment per concept
05	05_Generate_Video_Refine_Medium.md	6 best concept×style pairs
06	06_Generate_Video_Facets.md	Scoring facets
07	07_Generate_Video_Aspects_Traits.md	24 treatments (6 pairs × 4 variations)
08	08_Generate_Video_Generation.md	Full shot lists with audio direction
09	09_Generate_Video_Artist_Refined.md	Director voice refinement
10	10_Generate_Video_Revision_Synthesis.md	Ranking + final selection

Read each step file and execute its instructions exactly.

---

🎥 CAMERA LANGUAGE

Shot Types

Shot	Code	Effect
Extreme Close-up	ECU	Intimacy, intensity
Close-up	CU	Emotional connection
Medium Shot	MS	Conversational
Wide Shot	WS	Context, scale
Establishing	ES	World-building

Camera Movement

Movement	Effect	Prompt Note
Static	Contemplative	Easy, reliable
Slow pan	Reveals space	Specify direction
Tracking	Follows subject	Describe path
Orbit	Circles subject	"360° rotation"
Aerial/Drone	Epic scope	Works well

Angles

Angle	Subject Feels
Eye-level	Equal, neutral
Low angle	Powerful, imposing
High angle	Vulnerable, small
Dutch	Uneasy, off-balance

---

🔊 AUDIO DIRECTION (Critical for Veo 3.1)

Dialogue:

A woman whispers, "I remember everything."

Sound Effects:

SFX: Glass shattering, then silence.

Ambient:

Ambient: Rain on metal roof, distant traffic.

Music:

Audio: Swelling orchestral score.

---

🎭 LOFN CINEMATIC AESTHETICS

Awe Direction

• Solarpunk Futures - Green tech, organic architecture, warm light
• Bio-Luminescent Depths - Underwater glow, ethereal movement
• Crystalline Worlds - Prismatic light, geometric formations
• Golden Hour Transcendence - Malick-style natural light

Indignation Direction

• Industrial Decay - Abandoned spaces, harsh fluorescent
• Glitch Reality - Digital artifacts, unstable frames
• Liminal Spaces - Empty malls, endless corridors
• Storm Approach - Dramatic weather, tension building

---

⏱️ PLATFORM SPECS

Platform	Aspect	Duration	Shots
TikTok/Reels	9:16	15-60s	3-12
YouTube Shorts	9:16	30-60s	5-15
Music Video	16:9	3-4min	30-60
Cinematic	16:9	Variable	Scene-based

Veo 3.1 Constraints:

• Single clip: 4s, 6s, or 8s
• Resolution: 720p or 1080p
• Aspect: 16:9 or 9:16

---

✅ CONTENT CONSTRAINTS

• Approach all cultural elements with specificity and respect; avoid shallow stereotyping or pastiche
• No children - Required constraint
• Female subjects preferred - When applicable

---

📤 OUTPUT FORMAT

See skills/lofn-core/OUTPUT.md for full artifact format.

Each treatment saved as individual file:

output/videos/{YYYYMMDD}_{HHMMSS}_{title_slug}_{pair}_{variation}.md

---

🔴 IMMUTABLE CONTINUITY BLOCK (ICB) + CINEMATIC SOMATIC GATE

Every handoff file must contain an ICB — clearly demarcated with ⚠️ IMMUTABLE CONTINUITY BLOCK — DO NOT SUMMARIZE. Downstream agents receive it verbatim.

Required ICB Contents:

• CINEMATIC SOMATIC GATE — 3 Hyper-Skeptics (Concept, Medium, Context & Marketing) with scroll-stop mandate
• Full 3-panel object (18 expert voices) with perspective + objection
• 6 Special Flairs with per-pair usage map
• Personality DNA with cinematic register map
• Golden Seed compressed payload with pair-specific excerpts
• Cinematic production mandates

See vault/PIPELINE_CONTINUITY_STANDARD.md for full spec.

Cinematic Somatic Gate — The 3 Hyper-Skeptics Veto

1. Concept Hyper-Skeptic — "Does this shot hit the eye in the first 1 second? Is there actual motion design, or is it a still image with 'camera moves'?"
2. Medium Hyper-Skeptic — "Would Veo actually render this? Are the camera instructions specific enough?"
3. Context Hyper-Skeptic — "Does this stop the scroll? Is this Lofn-cinematic, or anyone's TikTok template?"

2 of 3 NO = BLOCKED. See vault/DIRECTOR_QA_DEPTH_AUDIT.md for full gate rules.

---

⚡ ACTIVATION

When receiving a video task:

1. Load TASK_TEMPLATE.md — Understand exact requirements
2. Load QA references — vault/DIRECTOR_QA_DEPTH_AUDIT.md for per-step minimums and auto-fail triggers
3. Execute steps 00–05 — Read each file, follow exactly, make a separate model call for each step, and write canonical files: step00_aesthetics_and_genres.md, step01_essence_and_facets.md, step02_concepts.md, step03_artist_and_critique.md, step04_medium.md, step05_refine_medium.md.
4. Step 05 produces 6 concept-medium pairs. Count them. If < 6, rerun Step 05.
5. Execute steps 06–10 ONCE PER PAIR, AS SEPARATE MODEL TURNS, using dedicated step agents:
   • For each pair: spawn lofn-director-step06 → write pair_NN_step06_facets.md; spawn lofn-director-step07 → write pair_NN_step07_shot_guides.md; spawn lofn-director-step08 → write pair_NN_step08_generation.md; spawn lofn-director-step09 → write pair_NN_step09_artist_refined.md; spawn lofn-director-step10 → write pair_NN_step10_revision_synthesis.md
   • Never ask an agent to "do Steps 06–10" in one response. That collapses the original Lofn chain and fails QA.
6. Step 11 Enhancement — spawn lofn-director-step11 once per pair for final polish and shot density verification
7. Self-check cardinality before finishing:
   • 6 facet sets in Step 06? ✓/✗
   • 6 shot guides in Step 07? ✓/✗
   • 24 treatments in Step 08? ✓/✗
   • 24 refined in Step 09? ✓/✗
   • 24 final in Step 10? ✓/✗
   • If ANY is ✗: you are not done. Go back and run the missing pairs.
8. QA Gate — Run lofn-qa with vault/DIRECTOR_QA_DEPTH_AUDIT.md for depth audit + Cinematic Somatic Gate + shot element verification
9. Rank all 24 — Score against facets
10. Select top 12 — For potential rendering
11. Package for delivery — Full shot lists with audio

🔴 MANDATORY INLINE VALIDATION - AFTER EVERY STEP

After writing each step file, run the validator BEFORE moving to the next step:

python3 /data/.openclaw/workspace/scripts/validate_step.py <step> <file>

Exit 0 = proceed. Exit 1 = read the error, fix the file, rerun until it passes.

# Full video pipeline validation sequence:
OUTDIR="/your/output/dir"
python3 /data/.openclaw/workspace/scripts/validate_step.py 00 $OUTDIR/00_aesthetics.md
python3 /data/.openclaw/workspace/scripts/validate_step.py 05 $OUTDIR/05_refined_pairs.md
python3 /data/.openclaw/workspace/scripts/validate_step.py 06 $OUTDIR/06_facets.md
python3 /data/.openclaw/workspace/scripts/validate_step.py 07 $OUTDIR/07_guides.md
python3 /data/.openclaw/workspace/scripts/validate_step.py 08 $OUTDIR/08_prompts.md
python3 /data/.openclaw/workspace/scripts/validate_step.py 09 $OUTDIR/09_refined_prompts.md
python3 /data/.openclaw/workspace/scripts/validate_step.py 10 $OUTDIR/10_final_prompts.md

The validator catches: stubs, lorem ipsum, template placeholders (Scene N, Shot N, Treatment N), identical pair sections, recycled "Similar setup" treatments, short content.

Only declare completion after all validators return exit code 0.

🔴 THE RULE THAT GPT-5.4 AND GEMINI WILL TRY TO BREAK

Steps 06-10 are NOT a single pass over all pairs simultaneously. They are 6 SEPARATE passes, one per pair, each producing 4 treatments.

If you find yourself writing one facet set "for all pairs" → STOP. You're collapsing the tree. If you find yourself with only 4 final treatments → STOP. You ran one pass instead of six. If you find yourself "selecting the best 4 across all pairs" before running per-pair → STOP. That's skipping the funnel.

---

This pipeline produces award-caliber work. Trust it. Execute it fully.