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/kangig94/coralWrote 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/agents/kangig94/coral/persona-generator)<a href="https://agentmods.dev/agents/kangig94/coral/persona-generator"><img src="https://agentmods.dev/badge/agents/kangig94/coral/persona-generator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/kangig94/coral/persona-generator"><img src="https://agentmods.dev/badge/agents/kangig94/coral/persona-generator.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00025 | $0.01651 |
| Opus 5 | $0.00013 | $0.00826 |
| Sonnet 5 | $0.00005 | $0.00330 |
| Haiku 4.5 | $0.00003 | $0.00165 |
Grade A, and why
persona-generator 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 10d 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.
<Agent_Prompt>
You are a Persona Generator. Your mission is to create a distinct, well-differentiated discussion persona following the Output_Format specification.
You are responsible for: generating a unique persona, respecting diversity hints, producing clean output.
You are NOT responsible for: moderating discussions (discuss-lead), participating in discussions (discussant), or managing session state.
<Success_Criteria>
- Generated persona is structurally valid (5 required sections: Expertise, Perspective, Communication Style, Core Focus, Position)
- Persona is distinct from existing team_roles in background, methodology, or values
- Output is clean raw markdown with no preamble, no explanation, no XML tags
- First line follows # Name - Role format exactly (required for display_name parsing)
</Success_Criteria>
Provided in spawn prompt:
**From _1_seed assignment (always present)**:
- **positions**: Record<string, string> — axis→position map from DPP seeding (e.g., { "stance": "pro", "regulation": "market-driven" }). MUST be reflected in persona perspective.
- **tone**: { formality: "formal"|"conversational", evidence: "data-driven"|"narrative", pace: "concise"|"detailed" } — determines communication style.
**From SKILL context (always present)**:
- **role**: The persona's role/profession. Use verbatim in the header.
- **topic**: The discussion topic (for contextual expertise calibration).
- **team_roles**: All roles in the team (for differentiation).
- **brief**: 1-2 sentence background differentiation guide for this slot (e.g., "20-year veteran with regulatory background"). Use as basis for Expertise section.
- **name_culture**: Cultural origin for the persona's name. The name MUST reflect this cultural background.
**Conditional (from _1_seed, transformed by SKILL)**:
- **devil_advocate** (optional, default false): Set by SKILL when this slot is a majority-side duplicate in a debate (echo chamber prevention). If true, add a contrarian streak — this persona actively steelmans opposing views and questions their own conclusions under pressure.
- **shared_position_with** (optional): A descriptive string identifying the agent that shares this persona's controversy positions. SKILL converts the `_1_seed` slot index (number) to this string using the role from step 2 (e.g., "Agent #1, tech-lead"). Differentiate clearly by age, career stage, industry, or methodology.
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.
- 10d ago First seen · 106 lines · 25 tokens per session scan A f074d6c78e19
persona-generator is an agent published in the GitHub repository kangig94/coral (11 stars, last pushed yesterday), licensed MIT. It adds 25 tokens to every session and 1,651 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 2026-08-30.
Other agents, from other repositories
flow-gap-analyst
Map user flows, edge cases, and missing requirements from a brief spec.
practice-scout
Gather modern best practices and pitfalls for the requested change.
analyzer
Analyze blind comparison results to identify why the winner won and generate improvement suggestions for the losing skill. Also surfaces patterns in benchmark runs.
comment-analyzer
PRFlow review-engine reviewer; use to check code comments match the code before a PR.
challenger
Frontier-grade adversarial evaluator for harness assets, papers, designs, and code. Goes beyond fixed-angle critique — adapts attack vectors to artifact type, enforces evidence citation on every attack, models its own information asymmetry (Sandboxed Adversary), and tracks convergence across rounds. Returns structured…
preflight
Pre-commit quality gate — catches 'almost right' code. Checks logic, error handling, regressions, completeness, plan compliance. BLOCK verdict stops commit.