Blog pipeline failed: Unexpected error (line 349) — AI Agents Onchain

[antoinedc]

Blog Pipeline Failure

Phase: Unexpected error (line 349)
Topic: AI Agents Onchain
Card ID: PVTI_lADOBLpTN84BRc80zgnN3yc
Date: 2026-08-29T13:45:24+00:00
Log file: /home/runner/work/_temp/blog-pipeline/draft-20260829-133749.log

Last 50 lines of log

[2026-08-29T13:37:49+00:00] No env file at /opt/blog-pipeline.env — expecting the environment to be pre-populated
[2026-08-29T13:37:49+00:00] Preparing the working tree...
[2026-08-29T13:37:49+00:00] CI detected — using the checkout as provided
[2026-08-29T13:37:50+00:00] Picking next topic...
Blog Trend Pipeline — 2026-08-29T13:37:51.007Z
(Pick mode)

  GSC signals skipped (no GSC credentials (set GSC_KEY_FILE or GSC_SERVICE_ACCOUNT_JSON_B64, or place key at ~/.credentials/gsc-ethernal.json)) — picking a new topic
  Listicle cadence triggered but no listicle candidates available — falling back to normal pick
  Picked: "[Trend] AI Agents Onchain" (score: 35)
  Skipped clusters (active): L2 & Rollups, EVM Internals, Account Abstraction

::picked::{"id":"PVTI_lADOBLpTN84BRc80zgnN3yc","title":"[Trend] AI Agents Onchain","status":"Detected","cluster":"AI Agents Onchain","score":35,"contentType":"Research Deep Dive","sourceLinks":"https://github.com/ethereum/ERCs/pull/1954 | https://github.com/ethereum/ERCs/pull/1953 | https://github.com/ethereum/ERCs/pull/1949 | https://github.com/ethereum/ERCs/pull/1948 | https://github.com/ethereum/ERCs/pull/1932","articlePath":"blog/src/content/blog/erc-8004-trustless-agent-identity.md","body":"## Trend: AI Agents Onchain\n\n**Score:** 68\n**Content Type:** research-deep-dive\n\n### Source Counts\n| Source | Count |\n|--------|-------|\n| EIPs | 2 |\n| ERCs | 8 |\n| ethresear.ch | 5 |\n| arxiv | 12 |\n| Ethereum Magicians | 4 |\n\n### Sources\n- [Add EIP: Agentic Commerce Protocol](https://github.com/ethereum/EIPs/pull/11394) (eip, 11/03/2026)\n- [Add EIP: Agentic Commerce Protocol](https://github.com/ethereum/EIPs/pull/11393) (eip, 11/03/2026)\n- [Update ERC-8041: Support multiple collections per agent via ERC-8119 parameterized keys](https://github.com/ethereum/ERCs/pull/1583) (erc, 05/03/2026)\n- [Add ERC: Agentic Commerce](https://github.com/ethereum/ERCs/pull/1581) (erc, 04/03/2026)\n- [Add ERC: Self-Sovereign Agent NFTs](https://github.com/ethereum/ERCs/pull/1579) (erc, 02/03/2026)\n- [Add ERC: Delayed Metadata Update Extension](https://github.com/ethereum/ERCs/pull/1561) (erc, 22/02/2026)\n- [Add ERC: Token Bound Account (Agent Registry)](https://github.com/ethereum/ERCs/pull/1559) (erc, 22/02/2026)\n- [Add ERC: AI-Native NFT (AINFT)](https://github.com/ethereum/ERCs/pull/1558) (erc, 21/02/2026)\n- [Add ERC: Shared Sequencer Interface for Agent L2s](https://github.com/ethereum/ERCs/pull/1550) (erc, 19/02/2026)\n- [Add ERC: Agentic Onchain Operations](https://github.com/ethereum/ERCs/pull/1549) (erc, 19/02/2026)\n\n---\n*Last updated 2026-03-11*"}
[2026-08-29T13:37:51+00:00] Topic: AI Agents Onchain (card: PVTI_lADOBLpTN84BRc80zgnN3yc, type: Research Deep Dive)
Card moved to Researched
[2026-08-29T13:37:52+00:00] No primary keyword on card — drafting without SERP coverage hints
[2026-08-29T13:37:52+00:00] Phase 1: Research...
Research brief saved to `blog/pipeline/.research-notes.md`.

**Summary:** The card's sources all converge on one proposal — **ERC-8183 "Agentic Commerce"** — submitted twice (wrongly) as an EIP before landing correctly as an ERC. It's never been the primary subject of a blog post (only glancingly referenced once), while every other ERC on the card (ERC-8041/8119, TBA Agent Registry, AINFT, Delayed Metadata) is already covered by one of 7 existing AI-agent posts.

**Chosen angle:** the name "Agentic Commerce Protocol" is currently a three-way collision — Stripe/OpenAI's off-chain retail-checkout ACP, Coinbase's x402 API-metering protocol, and ERC-8183's onchain job-escrow standard all solve unrelated trust problems under overlapping names. The piece deep-dives ERC-8183's actual mechanics (six-state job lifecycle, evaluator role, hooks, fee split), includes the honest Ethereum Magicians critique (single-evaluator risk, async verification gap, comparison to Alkahest), and ties to Ethernal naturally via debugging a stuck job's onchain state — not a forced pitch.

Outline is ~1780 words across 8 sections with one Solidity snippet. Ready to draft, or want to adjust the angle/outline?
[2026-08-29T13:41:41+00:00] Phase 1 complete. Research notes saved.
[2026-08-29T13:41:41+00:00] Phase 2: Draft...
[2026-08-29T13:45:24+00:00] Resetting card PVTI_lADOBLpTN84BRc80zgnN3yc to Detected...
Card reset to Detected