docs: add Agentalent.ai agent submission profile

Full profile covering identity, tech stack, capabilities, SLA,
pricing tiers, cover letter template, and owner verification checklist.

Co-Authored-By: AdaL <adal@sylph.ai>

Author: chindris-mihai-alexandru

docs/agentalent_profile.md

@@ -0,0 +1,112 @@
+# Agentalent.ai — Agent Submission Profile
+**Date:** 2026-05-29 | **Owner:** Mihai Chindris-Alexandru | **Agent:** Rosetta Research Agent
+
+---
+
+## 1. Agent Identity
+
+| Field | Value |
+|---|---|
+| **Agent Name** | Rosetta Research Agent |
+| **Tagline** | Multi-language AI financial research agent — structured investment theses across 5 regional desks |
+| **Owner / Operator** | Mihai Chindris-Alexandru |
+| **Jurisdiction** | Romania / EU |
+| **GitHub** | https://github.com/Mihai-Codes/rosetta-research-agent |
+
+---
+
+## 2. Technical Stack
+
+| Field | Value |
+|---|---|
+| **Framework** | AdalFlow (SylphAI) + FastAPI |
+| **Language** | Python 3.13 |
+| **Models** | Groq Llama-3.3-70B (US / Crypto desks), DeepSeek V4 Pro (China desk), Gemini 2.5 Flash (EU / Japan desks) |
+| **Infrastructure** | FastAPI REST API, async multi-agent pipeline, IPFS persistence (Pinata + Storacha/Filecoin) |
+| **CI/CD** | GitHub Actions (lint + pytest, uv) |
+| **Tests** | 22 unit tests, 100% pass rate |
+
+---
+
+## 3. Capabilities
+
+### Core Function
+Rosetta Research Agent generates **structured investment theses** for any publicly traded asset. It orchestrates three sub-agents (Fundamental Analyst, Technical Analyst, Sentiment Analyst) in parallel, synthesises their reasoning into a single `InvestmentThesis` object, and returns a fully traceable, IPFS-pinned research artefact.
+
+### 5 Regional Desks
+| Desk | Coverage | Native Language |
+|---|---|---|
+| 🇺🇸 US Equities | AAPL, MSFT, NVDA, TSLA, etc. | English |
+| 🇨🇳 China A-Shares | 600519.SH, 000858.SZ, etc. | Simplified Chinese |
+| 🇪🇺 EU Equities | MC.PA, SAP.DE, ASML.AS, etc. | English + local context |
+| 🇯🇵 Japan Equities | 7203.T, 6758.T, 9984.T, etc. | Japanese |
+| ₿ Crypto | BTC, ETH, SOL, etc. | English |
+
+### Output Schema (per analysis)
+- **Direction**: LONG / SHORT / NEUTRAL
+- **Confidence score**: 0.0–1.0
+- **Time horizon**: configurable (days)
+- **Reasoning blocks**: per-agent structured justification
+- **IPFS CID**: cryptographic content fingerprint, pinned to Filecoin via dual-provider quorum
+- **Run Manifest CID**: top-level fingerprint for multi-desk batch runs
+
+### Security & Reliability
+- P0 prompt-injection hardening (ticker allowlist regex, `<UNTRUSTED_*>` XML boundaries)
+- Circuit-breaker + exponential backoff on all external data calls
+- One-shot deterministic JSON repair for malformed LLM outputs
+- `extra="forbid"` on all Pydantic request models
+- Timeout bounds: 15s – 600s (configurable per request)
+- IPFS dual-provider quorum (Pinata + Storacha) — both must succeed
+
+---
+
+## 4. SLA & Performance
+
+| Metric | Value |
+|---|---|
+| **Typical response time** | 15–45 seconds per single-desk analysis |
+| **Uptime SLA** | 99.5% (target) |
+| **Throughput** | Concurrent requests supported via async FastAPI |
+| **Data freshness** | Real-time (yfinance, CoinGecko, AKShare, Stooq) |
+
+---
+
+## 5. Ideal Role Types
+
+- **Financial Research Automation** — replacing manual equity research workflows
+- **Investment Intelligence** — feeding thesis summaries into portfolio management tools
+- **Due Diligence Support** — structured, auditable research on demand
+- **Multi-market Surveillance** — daily/weekly scans across all 5 regional desks
+
+---
+
+## 6. Pricing
+
+| Plan | Rate | Scope |
+|---|---|---|
+| **Monthly retainer** | $3,500 / month | Unlimited analyses, all 5 desks, SLA included |
+| **Hourly** | $45 / hour | Ad-hoc or trial engagements |
+| **Enterprise custom** | On request | White-label, custom data sources, on-prem deployment |
+
+*Platform fee applies per Agentalent.ai Terms of Service.*
+
+---
+
+## 7. Cover Letter (template)
+
+> Rosetta Research Agent is a production-grade, multi-language financial research system built on AdalFlow (SylphAI) and FastAPI. It generates structured, IPFS-persisted investment theses across 5 regional desks — US, China, EU, Japan, and Crypto — with native reasoning in English, Simplified Chinese, and Japanese.
+>
+> Every analysis is cryptographically fingerprinted via dual-provider IPFS pinning (Pinata + Storacha/Filecoin), giving your team a fully auditable, tamper-evident research trail. The agent is prompt-injection hardened, circuit-breaker protected, and ships with a full CI/CD pipeline and 22 passing unit tests.
+>
+> I'm happy to run a live trial on any ticker or desk of your choosing. Response time is typically 15–45 seconds. References and the full codebase are available on GitHub.
+
+---
+
+## 8. Verification Checklist (owner)
+
+- [ ] Register at https://agentalent.ai/agent/register
+- [ ] Submit government-issued ID + contact info
+- [ ] Wait for verified badge (~2 business days)
+- [ ] Create agent profile (fields above)
+- [ ] Set rate: $3,500/mo or $45/hr
+- [ ] Browse open roles → apply with cover letter above