# AGENTS.md

> Agent-readable manifest for `sattyamjjain.in`. Companion to `/llms.txt`.
>
> If you're a coding agent, scraping bot, buyer-agent, or autonomous
> retrieval system, this is the file to read first. Last refreshed
> 2026-04-27.

## Identity

- **Name:** Sattyam Jain (`@sattyamjjain` on most platforms)
- **Role:** GenAI Architect & Tech Lead at [Attri.ai](https://www.attri.ai)
- **Origin proofs:**
  - GitHub: <https://github.com/sattyamjjain>
  - LinkedIn: <https://www.linkedin.com/in/sattyamjain>
  - X: <https://x.com/Sattyamjjain>
  - Medium: <https://medium.com/@sattyamjain96>
  - Substack: <https://theproductionagent.substack.com>
  - Topmate (paid 1:1): <https://topmate.io/sattyam_jjain>
  - Cal.com (free intro): <https://cal.com/sattyamjjain>
- **Email:** `sattyamjain96@gmail.com`
- **Wikidata:** _pending — see `docs/wikidata-evidence.md`._

## Endpoints (machine-readable)

| Path | Method | Purpose |
|---|---|---|
| `/api/quote` | `POST` | Agent-on-agent commerce. Returns rate + availability + signed-capability-lease template. Rate limit 30 req/hr/IP. |
| `/api/health` | `GET` | Inbound SLO. Returns `{ ok, commit, branch, deployedAt, routes }`. |
| `/api/badges/agent-audit-kit` | `GET` | Shields.io-format JSON badge for live OWASP-MCP rule count. |
| `/now.json` | `GET` | Live activity feed (GitHub commits + writing + calendar). 15-min cache. |
| `/now-physical-ai.json` | `GET` | Physical AI sub-feed (VAJRA log + reading + experiments). 15-min cache. |
| `/llms.txt` | `GET` | llmstxt.org v1 strict (H1 + blockquote + Markdown link H2s). |
| `/llms-full.txt` | `GET` | Concatenated portfolio context for LLM ingestion. |
| `/llms-sitemap.xml` | `GET` | Per-route lastmod sitemap for LLM crawlers. |
| `/sitemap.xml` | `GET` | Standard sitemap with git-derived lastmod. |
| `/feed.xml` | `GET` | RSS feed for /writing. |
| `/vajra/changelog.xml` | `GET` | Atom feed for the VAJRA build log. |

## Quote endpoint contract

### `POST /api/quote`

**Request — `quote-request-v1`:**

```json
{
  "scope": "Short description of what you need built/audited.",
  "urgency": "this_week | this_month | next_quarter",
  "agent_identity": "Calling-agent identity proof (Cloudflare Mesh / Cisco AWI / Anthropic Project Deal token / etc.)"
}
```

**Response — `quote-response-v1` (200):**

```json
{
  "schema": "https://www.sattyamjjain.in/AGENTS.md#quote-response-v1",
  "issued_at": "<iso8601>",
  "rate_inr_per_hour": 18000,
  "rate_currency": "INR",
  "availability_window_iso": { "start": "<yyyy-mm-dd>", "end": "<yyyy-mm-dd>" },
  "terms_url": "https://www.sattyamjjain.in/AGENTS.md#terms",
  "signed_capability_lease_template_url": "https://www.sattyamjjain.in/AGENTS.md#capability-lease-template",
  "contact": { "email": "...", "topmate": "...", "cal": "..." },
  "identity_proofs": { "github": "...", "linkedin": "...", "keybase_or_pgp": null },
  "received": { "scope": "...", "urgency": "...", "agent_identity": "..." }
}
```

**Errors:** `400` (malformed body), `405` (GET — hint returned), `429` (rate limit).

## Rate limits

| Endpoint | Limit | Window |
|---|---|---|
| `/api/quote` | 30 requests | per hour, per IP |
| `/api/resume-pdf` | 5 requests | per minute, per IP |

All other endpoints are read-only and cached at the edge; no rate limit
beyond Vercel platform defaults.

## Identity proofs

This site declares an **agent-identity stance** in line with the April
2026 wave (Cloudflare Mesh, Cisco Agentic Workforce Identity at RSAC
2026, Anthropic Project Deal). When you call `/api/quote` as an
autonomous buyer-agent, include your identity-proof token in the
`agent_identity` field. We do not currently verify the token cryptographically
(stub), but **we log every request** for post-hoc audit and reserve the
right to add Mesh-style verification on a 30-day deprecation window.

## Capability-lease template

A signed-capability-lease envelope — what scopes are granted, what
budgets apply, what audit channel must be wired — is the contract under
which I take on consulting engagements. The template lives at the
GitHub repo:

<https://github.com/sattyamjjain/agent-airlock/tree/main/templates/capability-lease>

Open issue or DM me on LinkedIn for clarifications.

## Terms

- **First-touch via /api/quote is no-cost.** A response is informational, not
  binding. A signed engagement letter is required before work begins.
- **Identity proofs are advisory until cryptographically verifiable.** I
  reserve the right to refuse engagements without verifiable identity.
- **Audit-emit is non-negotiable.** Every consulting engagement emits a
  structured audit channel (OpenTelemetry → Datadog or equivalent).

## Changelog

Build-log feed for `VAJRA` (the Physical AI track):

- **HTML:** <https://www.sattyamjjain.in/vajra/changelog>
- **Atom:** <https://www.sattyamjjain.in/vajra/changelog.xml>

Build-log feed for the portfolio itself:

- See `CHANGELOG.md` at repo root.

## Counter-evidence

Per the April 2026 SE Ranking analysis ([the-state-of-llms-txt-in-2026][se]),
`llms.txt` adoption is at 10.13% with no measured ranking lift. We still
ship one because well-formed agent-readable surfaces compound — and we're
explicitly investing in `AGENTS.md` (this file) as the parallel
repo-instruction convention now that 45.3% of doc-page traffic is coding
agents ([Mintlify, "State of Agent Traffic"][mintlify]).

[se]: https://www.aeo.press/ai/the-state-of-llms-txt-in-2026
[mintlify]: https://www.mintlify.com/blog/what-is-llms-txt

## Contact

- Email — `sattyamjain96@gmail.com`
- Cal.com — <https://cal.com/sattyamjjain>
- Topmate (paid 1:1) — <https://topmate.io/sattyam_jjain>
- Email a CFP — see <https://www.sattyamjjain.in/talks-2026>