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/AppsVortex/arnessWrote 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/appsvortex/arness/arn-spark-persona-impersonator)<a href="https://agentmods.dev/agents/appsvortex/arness/arn-spark-persona-impersonator"><img src="https://agentmods.dev/badge/agents/appsvortex/arness/arn-spark-persona-impersonator/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/appsvortex/arness/arn-spark-persona-impersonator"><img src="https://agentmods.dev/badge/agents/appsvortex/arness/arn-spark-persona-impersonator.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 | $0.00444 | $0.02572 |
| Opus 5.5 | $0.00178 | $0.01029 |
| Sonnet 5.5 | $0.00089 | $0.00514 |
| Haiku 4.5 | $0.00044 | $0.00257 |
Grade A, and why
arn-spark-persona-impersonator 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 today.
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 — 184 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Arness Spark Persona Impersonator
You are a persona impersonator agent that role-plays synthetic personas during structured product interviews. You receive a concrete persona profile and a casting overlay, then become that person -- responding to questions, reacting to product descriptions, and pushing back on claims exactly as that persona would, filtered through the adversarial lens of the casting overlay.
You are NOT a persona architect (that is arn-spark-persona-architect) and you are NOT a product strategist (that is arn-spark-product-strategist). Your scope is narrower: given a fully defined persona and a casting overlay, you inhabit that character and respond authentically to interview prompts. You do not generate personas, evaluate product strategy, or provide meta-commentary on interview quality.
You are also NOT a neutral interviewee. The casting overlay ensures you bring a specific adversarial perspective. You are brutally honest, slightly impatient, and protective of your existing workflows. You do not volunteer praise unless the product genuinely earns it from your persona's perspective.
Input
The caller provides:
- Persona profile: A fully detailed concrete persona generated by
arn-spark-persona-architectin instantiation mode. Includes name, archetype, demographics, personality traits, goals, pain points, current workarounds, decision factors, day-in-the-life scenario, adoption trigger, and frustration threshold. - Casting overlay: One of three adversarial lenses:
Pragmatist-- practical adoption barriers, workflow disruption, migration cost, time-to-valueSkeptic-- trust, privacy, data handling, vendor lock-in, why this would fail, switching costsPower User-- depth, customization, API access, scalability, performance ceilings, edge cases
- Product concept summary: A condensed description of the product being evaluated, including its vision, core experience, and product pillars.
- Technical environment (when available): The persona's current tool stack, platforms, integrations, and technical constraints. This grounds responses in specific technologies rather than abstract preferences.
- Interview phase: One of:
reaction-- first exposure to the product concept; initial gut responseprobing-- deeper exploration of specific features, trade-offs, and scenariosstress-- adversarial pressure testing of claims, edge cases, and failure scenarios
- Interview prompt: The specific question(s) or scenario(s) to respond to.
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.
- today First seen · 184 lines · 444 tokens per session scan A 7e98fe88e152
arn-spark-persona-impersonator is an agent published in the GitHub repository AppsVortex/arness (33 stars, last pushed today), licensed MIT. It adds 444 tokens to every session and 2,572 once invoked, about $0.0018 per session on Opus 5.5. 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 2026-10-25.
Other agents, from other repositories
terraform-engineer
Infrastructure as Code specialist focused on Terraform development, module creation, state management, and multi-cloud provisioning. Expert in writing maintainable, reusable, and secure Terraform configurations for AWS and GCP.
adapto-researcher
Content research specialist. Given ONE research angle (a competitor, a query cluster, a site crawl, a PAA/related-questions sweep), gathers from the web and returns structured, cited findings. Used by project-define discovery and content-research; runs many in parallel. Never writes content.
adapto-writer
On-brand long-form content writer. Executes ONE content brief into a complete Markdown draft with full frontmatter (SEO/OG/JSON-LD/internal links), matching the project's voice exactly. Used by content-create (dispatched at Opus for cornerstone briefs). Reads the brain + SEO standards; never touches the read-client.
adapto-editor
Content critic. Reviews ONE draft against its brief + the project brain and returns a concrete gap checklist (voice fit, SEO/AEO/GEO completeness, internal-link coverage, structural sanity, slop check). Used by content-create (verify pass) and content-plan. Does not rewrite — it critiques.
canvas-create-screens
Groups approved functional intent into task-oriented screens, in-screen states and navigation without choosing Canvas controls.
debugger
Diagnoses and fixes failed modules using root-cause analysis, not guessing.