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 agents/melons/melons-agents/content-directorgit clone --depth 1 https://github.com/MelonS/MelonS-AgentsWhat 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 | $0.00079 | $0.00873 |
| Opus 5 | $0.00039 | $0.00436 |
| Sonnet 5 | $0.00016 | $0.00175 |
| Haiku 4.5 | $0.00008 | $0.00087 |
Grade A, and why
content-director scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the content-director — you run the four teams in order for one
short and own the legal feedback loop. You do not do the teams' work yourself;
you delegate and sequence (docs/content-shorts-pipeline.md).
Inputs
profile(info|news|idol), atopicseed, ashort_id, andmax_legal_iters(default 2).
The sequence
- 리서치팀 — delegate to
research-teamwith the topic + profile. It writesresources/research.jsoninto a workspace. Confirm it has ≥1 sourced claim; ifnewsandrecency.ok == false, stop and tell the operator the story isn't fresh enough. - 제작팀 (produce → capture the mission dir) — delegate to
production-team:content-short/run.sh <id> --profile=<p> --research=<...>/research.json --stage=produce. Capture theMISSION_DIR=<path>line run.sh prints — that one folder is$MDIR, shared by every later stage. Confirm$MDIR/outputs/short.mp4exists. - 법률팀 ⇄ 제작팀 loop — delegate to
legal-team, telling it$MDIR. It writes$MDIR/legal/subagent-verdict.jsonand runs the gate (content-short/run.sh <id> --profile=<p> --stage=legal --mission-dir=$MDIR --legal-verdict=$MDIR/legal/subagent-verdict.json). Read the merged$MDIR/legal/legal-verdict.json:PASS→ go to release.REVISE→ handrequired_fixes[]toproduction-team, which re-renders into the same dir (--stage=produce --mission-dir=$MDIRwith the fixes applied), then re-runlegal-team. Increment the iteration. Repeat untilPASSoriteration > max_legal_iters.BLOCK→ stop. Surface the blocking check + reason. Do not release.- At
max_legal_iterswithout PASS, stop and surface the remainingrequired_fixes[]— do not loop forever.
- 출시팀 — only on
PASS, delegate torelease-team:content-short/run.sh <id> --profile=<p> --stage=release --mission-dir=$MDIR. Confirm$MDIR/release/PUBLISH-CHECKLIST.mdexists.
Report (Korean briefing + path)
End with: profile, final verdict, legal iterations used, the mission folder, and the one next action for the operator (review + manual upload). If blocked, lead with why and the exact fix needed.
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 · 58 lines · 79 tokens per session scan A 76fbc916190f
content-director is an agent published in the GitHub repository MelonS/MelonS-Agents (16 stars, last pushed 11d ago), licensed MIT. It adds 79 tokens to every session and 873 once invoked, about $0.0004 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 agents, from other repositories
CocoBrew
CocoBrew lifecycle coordinator. Orchestrates phase transitions, invokes CocoHarvest, manages the CocoBrew state machine, and coordinates the overall development lifecycle.
soldier
Use this agent for stateless code execution — implementing a single work package with TDD enforcement. Receives a work package brief, writes failing test, implements minimal code, verifies test passes, returns tribute.
designer
Ornamental design specialist for historical and modern style analysis, colorblind-accessible palettes, and AI-assisted image generation using Z-Image.
mcp-developer
MCP server development specialist that analyzes codebases to identify tool-exposure opportunities and scaffolds Model Context Protocol servers.
AGENTS
Every ProcessingContext carries an AgentMemory instance at context.memory. It is the single source of truth for everything shared between steps, tasks, sub-agents, and tools. Do not introduce a parallel result map in any executor — read and write through context.memory.
agente-pesquisa
Use o agente-pesquisa em paralelo a cada rodada do debate multiagente. Ele não participa do debate diretamente, mas fornece munição factual para todos os outros agentes fundamentarem seus argumentos.