This is the record of an abcd-managed repository

abcdBetav0.16.2 · 2026-11-06

Development record

Alice Reads One Page and Knows Exactly Where abcd Stands in the Agentic-AI Landscape

intentshippedminorstandaloneitd-100

Press Release

Alice reads one page and knows exactly where abcd stands in the agentic-AI landscape. Alice evaluates developer tooling for their team. Every tool describes itself in its own invented vocabulary, and they have to reverse-engineer how it relates to the terms the industry actually uses — MCP, the agent loop, context engineering, guardrails, policy-as-code, evaluations. abcd gives them a single reference page: a terminology crosswalk mapping each established agentic-AI term to how abcd incorporates it (naming the native verb or principle), adapts it under a sharper name, deliberately rejects it for a recorded reason, or is still watching it via a ledger entry they can follow. Every established definition cites a primary source — a spec, a standards body, or a peer-reviewed paper — and every claim about abcd is verified against the committed record, never memory. "I stopped reverse-engineering the invented names," said Alice. "One page told me, in the vocabulary I already use, what this tool does — and, rarer still, what it deliberately does not."

Why This Matters

A tool that invents its own vocabulary asks every evaluator to do translation work it should have done itself, and the translation is where trust is won or lost: an unmapped term reads as either ignorance of the field or marketing evasion. The crosswalk makes the mapping a maintained, cited artefact instead of folklore — and it is deliberately honest in both directions, recording not only what abcd incorporates but what it rejects and why, with every rejection traceable to the committed record. A REJECTS row with a recorded reason is a stronger credibility signal than any feature claim.

Acceptance Criteria

  • Given Alice opens docs/reference/terminology.md, when they read any row, then it carries the term, a one-line established definition with a footnote citation, and abcd's position as exactly one of USES / ADAPTS / REJECTS / WATCHING.
  • Given any footnote is followed, then it resolves to a primary source — a specification, a standards body, a peer-reviewed paper with DOI, or the engineering document where the term originated — never an aggregator or secondary write-up.
  • Given any abcd-position claim in the table, then it is verifiable against a committed record artefact (a verb, principle, ADR, intent, or ledger id); WATCHING rows cite an open ledger entry by id.
  • Given abcd docs (docs-currency lint) runs over the repo, then the page passes, with host and vendor names confined to citation footnote lines under the sanctioned allow escape.
  • Given the page links to abcd-native concepts, then every cross-reference targets docs/** only; native terms carry an inline gloss rather than a deep link into the development record.

Open Questions

None recorded yet.

Audit Notes

Fidelity review — receipt rcp-8d75f3749de3 (verifier intent-fidelity-reviewer claude-fable-5).

Provenance: intent-fidelity-reviewer@claude-fable-5 · rubric_hash sha256:d1f6f6e7e9db1ceea44dd78279a564d2ced6b6904ffd7461dca652fbfac8042f · prompt_hash sha256:95792472ae74ca0469f69a51c618946e0d33cb1380032460099ed4b469d67e86 Input attestations: diff:3f0ff9b..2998402 (PR #148, merge 2998402)@sha256:fec2fe2c0c7203cf07d5039ebab4f61245fe341ea58332946b2b7cdd0a8cc217; rubric:.abcd/.work.local/reviews/rcp-8d75f3749de3.request.md@sha256:d1f6f6e7e9db1ceea44dd78279a564d2ced6b6904ffd7461dca652fbfac8042f;

Acceptance rollup: MET 4 · MET_WITH_CONCERNS 1 · NOT_MET 0 · INCONCLUSIVE 0

Per-criterion verdicts:

  • ac-1 — MET: grep-verified: 26 table rows, all 26 carry a footnote marker, all 26 open the position column with exactly one of the four labels (8 USES / 10 ADAPTS / 7 WATCHING / 1 REJECTS, matching spc-15's declared distribution); zero rows carry two labels or none evidence: docs/reference/terminology.md:40 — "| Term | Established meaning | abcd's position |" evidence: docs/reference/terminology.md:42 — "| Agent harness | The infrastructure around the model …[^harness] | USES —" evidence: .abcd/development/specs/closed/spc-15-alice-reads-one-page-and-knows-exactly.md:29 — "26 rows (8 USES / 10 ADAPTS / 7 WATCHING / 1 REJECTS)"
  • ac-2 — MET: every footnote (lines 71-121) cites a specification, standards body, DOI-bearing paper, or origin engineering doc — no aggregators; the delivered diff includes a fetch-validation commit (52 URLs + 11 DOIs on this page, zero broken), and my independent spot-check of the least familiar URL (learn.chatgpt.com/docs/sandboxing) resolved to the genuine primary sandboxing doc evidence: docs/reference/terminology.md:119 — "IETF, RFC 9162, "Certificate Transparency Version 2.0", December 2021, DOI 10.17487/RFC9162" evidence: docs/reference/terminology.md:115 — "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks", NeurIPS 2020, DOI 10.48550/arXiv.2005.11401" evidence: commit f6d8b22 (docs: correct citation drift found by the validation sweep) — "fetched all 84 citation elements across the docs (52 URLs and 11 DOIs on the terminology page alone): zero broken links"
  • ac-3 — MET_WITH_CONCERNS: all 21 cited record ids resolve to committed artefacts with matching lifecycle state (itd-3 shipped, itd-29 planned, itd-33/16/17/18/51 drafts, itd-5/81 disciplines, adr-23/24/25/27/29/35, iss-26/62/138–141 all in issues/open/) and content spot-checks match the claims; CONCERN: three of seven WATCHING rows (A2A, durable execution, observability) cite draft/planned intent or ADR ids but no open issues-ledger iss-N id — a divergence from the criterion's letter that spc-15 records as the signed-off widening to "an open record id" evidence: .abcd/development/decisions/adrs/0035-lifeboat-as-coverage-experiment.md:94 — "The graveyard's cite-or-be-dropped validator generalises" evidence: .abcd/development/intents/disciplines/itd-81-judge-calibration.md:135 — "known-good cases are ≥40% of the corpus" evidence: .abcd/development/brief/05-internals/08-skills.md:3 — "abcd ships zero skills" evidence: .abcd/work/issues/open/iss-138-payments-protocols-no-position.md:2 — "id: "iss-138"" evidence: docs/reference/terminology.md:47 — "WATCHING — a draft record (itd-33) leans against it" evidence: .abcd/development/specs/closed/spc-15-alice-reads-one-page-and-knows-exactly.md:110 — "WATCHING rows name an open record id (itd-16, itd-29, itd-33, iss-62"
  • ac-4 — MET: abcd docs lint run 2026-07-27 at HEAD ab8876a exits 0 with 0 blockers (the single WARN is in docs/reference/cli/commands.md:308, a different file) and terminology.md is byte-identical to the merged state (git log 2998402..HEAD on the file is empty); a vendor-token grep over the page body above the Sources section finds none, and each footnote line carrying a token the line-scoped harness/* family bans (Codex, Gemini, claude-domain URLs) carries the sanctioned allow escape evidence: abcd docs lint (run output) — "abcd docs lint — 1 finding(s), 0 blocker(s)" evidence: .abcd/docs-lint.json:14 — ""id":"harness/claude-code"…"allow_context":["(?i)< !--\\sdocs-lint:\\sallow\\b"]" evidence: docs/reference/terminology.md:71 — "https://www.anthropic.com/engineering/effective-harnesses-for-long-running-agents. < !-- docs-lint: allow -- >"
  • ac-5 — MET: grep for markdown links finds zero in-repo links anywhere in terminology.md — record ids appear only as plain text with an explicit inline-gloss statement, so no cross-reference leaves docs/; the one inbound link lives in docs/reference/README.md within docs/ evidence: docs/reference/terminology.md:19 — "Identifiers such as itd-N, iss-N, adr-N, and spc-N refer to entries in abcd's development record" evidence: docs/reference/terminology.md:21 — "native terms are glossed inline rather than deep-linked" evidence: docs/reference/README.md:7 — "terminology.md (linked) — the terminology crosswalk"

Gap audit:

  • honoured:
    • a single reference page mapping each established term to USES / ADAPTS / REJECTS / WATCHING evidence: docs/reference/terminology.md:40 — "| Term | Established meaning | abcd's position |"
    • every established definition cites a primary source, verified rather than assumed evidence: commit f6d8b22 (commit message) — "zero broken links, every DOI resolves to the cited work"
    • honest in both directions — a REJECTS with a recorded reason, and a payments silence captured as iss-138 instead of an invented rejection evidence: docs/reference/terminology.md:45 — "captured as iss-138 rather than dressed up as a rejection" evidence: .abcd/development/brief/05-internals/08-skills.md:3 — "abcd ships zero skills"
    • every claim about abcd is verified against the committed record, never memory evidence: dcn-2607260000000001 (delivered grill line, 2026-07-26) — "AP2 via new capture — record silent, REJECTS never invented"
    • the page ships wired into the user-facing surface with a CHANGELOG entry and an inspiration credit landed in the same change evidence: docs/reference/README.md:7 — "terminology.md (linked)" evidence: CHANGELOG.md (delivered hunk) — "A public terminology crosswalk at docs/reference/terminology.md" evidence: ACKNOWLEDGEMENTS.md (delivered hunk) — "Credited as the prompt, not a source"
  • diverged:
    • press release: watching positions are followable "via a ledger entry" — delivered: three of seven WATCHING rows (A2A, durable execution, observability) track via draft/planned intent or ADR ids with no iss-N ledger entry, per the spec's recorded widening to "open record id" evidence: docs/reference/terminology.md:52 — "WATCHING — planned run resilience (itd-29)" evidence: .abcd/development/specs/closed/spc-15-alice-reads-one-page-and-knows-exactly.md:110 — "WATCHING rows name an open record id"
    • press release: primary sources are "a spec, a standards body, or a peer-reviewed paper" — delivered admission rule additionally accepts origin vendor engineering docs (consistent with ac-2's fuller wording, and recorded as spec design decision 3) evidence: docs/reference/terminology.md:17 — "or engineering documentation from the vendors where the term originated" evidence: .abcd/development/specs/closed/spc-15-alice-reads-one-page-and-knows-exactly.md:70 — "vendor engineering posts admissible only where they are the origin of the term"
  • missing:
    • spc-15's conditional promise that deferring the structural docs-lint rule ("every table row carries at least one footnote") would be "recorded, not silent" — no such rule ships in the diff and no implement-time deferral record appears in the delivered DECISIONS.md line or ledger captures evidence: .abcd/development/specs/closed/spc-15-alice-reads-one-page-and-knows-exactly.md:46 — "If deferred, that is recorded, not silent" evidence: dcn-2607260000000001 (delivered grill line, 2026-07-26) — "no mention of the structural lint-rule deferral"

Frontmatter

iditd-100
typeintent
lifecycleshipped
kindstandalone
severityminor

.abcd/development/intents/shipped/itd-100-alice-reads-one-page-and-knows-exactly.md
open on GitHub ↗ · commit history ↗

Referenced by3

  • itd-2609061543533170One verb takes a managed repository's site from the checkout to a live addressrelated
  • itd-2609212103568351The status board shows Now, Next and Later, computed from the recordrelated
  • spc-15alice-reads-one-page-and-knows-exactly-where-abcd-stands-inimplemented by

mentions18

  • adr-23A transport-agnostic Go core behind thin front doors
  • dcn-26072600000000012026-07-26 — itd-100 grill settled: crosswalk is a mapping not a registry (no canonical native definitions; gloss-only rows, docs/** links only, pending iss-40); alphabetical + thematic mini-index; footnote citations with per-line docs-lint allow for vendor names; British English. Positions: A2A and AP2/payments both WATCHING (AP2 via new capture — record silent, REJECTS never invented; x402 + OpenAI/Stripe ACP folded into the AP2 footnote); policy engine folded into policy-as-code row; agent skills is the sole REJECTS (brief 08-skills commands-only rule). Rejected-term admissions blessed (agentic workflow, ANP, AGNTCY, Verifiable Intent, harness engineering, ISO 22989-as-HITL-anchor). LinkedIn anecdote omitted from itd-100. Two OpenAI harness URLs await maintainer read before ship.
  • dcn-26072700000000022026-07-27 — DEFERRED the row-has-footnote structural docs-lint rule (spc-15 out-of-scope): no existing rule shape covers table-row-to-footnote structure, so it would need a bespoke check; the grill left it optional and the citation-gate intent (itd-101) is now its natural home — implement it there or not at all. Recorded per spc-15's deferral-is-recorded clause; surfaced by the itd-100 fidelity audit's gap check.
  • iss-26sources corpus retrieval seam: the MVP is the current native store (per-source folders, text.md extraction, grep-based consult — shipped) and retrieval/processing becomes a pluggable seam with opt-in deeper backends, mirroring the spec seam pattern (native minimal default, adapter for depth): a RAG backend (RAG-Anything, already downloaded locally, has its own content processor that could replace pandoc/pdftotext extraction and serve semantic retrieval) and/or a CMS backend. Adapter contract, not a rewrite: consult and ingest skills keep one surface; the backend is configuration.
  • iss-62abcd-managed repos must guarantee the user's chosen git identity (name + email) is honoured on every commit. ahoy doctor should detect when the effective git author/committer identity diverges from the user's GitHub identity — a stray repo-local user.name/user.email override (e.g. a sandbox 'Test User <test@example.com>' shadowing the global identity), or committer != author — and, rather than silently setting it, PROPOSE a default (the global git identity, or the gh-authenticated account) and ASK the user to confirm before writing repo-local config. Motivation: a repo-local 'Test User <test@example.com>' override silently authored/committed 54 commits (pushed to origin) before detection this run, forcing a history rewrite + force-push to correct. A managed-repo identity gate at ahoy/doctor time catches it before the first commit.
  • iss-138payments protocols have no recorded position: the itd-100 terminology crosswalk found the record silent on agent-led payments (AP2 at the FIDO Alliance, x402, the two-vendor Agentic Commerce Protocol). The crosswalk row is WATCHING because a REJECTS needs a recorded reason; this capture is where that reason gets decided — likely out-of-scope-by-design for a development-configuration tool, but that is a maintainer call, not an inference. Detector: the crosswalk row cites this id; resolving it means recording the position and updating the row.
  • iss-139SKILL.md compatibility watch: the record rules abcd ships zero skills (commands-only namespace; any state mutation is a command). The external Agent Skills spec (agentskills.io) is unversioned, draft-stage, and without a standards-body steward, but multi-vendor adoption is real. If a findings-only skill ever ships, SKILL.md is the accepted executable form — this capture tracks the spec's maturation (versioning, stewardship, the Experimental fields) so that decision is made against current facts. The itd-100 crosswalk's REJECTS row cites this id.
  • iss-140AGENTS.md stewardship citation watch: abcd's AGENTS.md router predates the format's contribution to the Agentic AI Foundation (Linux Foundation, December 2025). The standardisation strengthens the bet, but the record nowhere cites the AAIF stewardship; the itd-100 crosswalk row now does so from docs/. Resolving this means deciding whether the brief's install/router surfaces should cite the external standard, and watching for spec versioning under the foundation.
  • iss-141tamper-evident receipts: receipts and review directories are hash-anchored and manifests are verified, but nothing is cryptographically signed and there is no append-only chain over the receipt history — integrity evidence, not proof against a capable insider. itd-16 (hash-chain/Merkle audit umbrella, draft) is the design home; this capture tracks the narrower question of whether the existing receipt flow should become tamper-evident (RFC 9162-style verifiable log or signing) independent of the full audit umbrella. The itd-100 crosswalk's tamper-evidence row cites this id.
  • iss-142prompt-quality brief self-contradiction: 05-prompt-quality.md states PQ001-PQ006 are shipped by spc-8, while 06-lint.md declares the PQ family a design target with no PromptLinter in internal/core/lint — the code sides with 06-lint. An instance of the iss-37 phantom-enforcement class, surfaced by the itd-100 docs-currency review when the crosswalk nearly repeated the phantom claim. Detector: enforcement-claims-are-facts sweep over the brief's Shipped-by annotations; acceptance corpus: the two contradicting lines.
  • iss-321persona pronoun discipline itd-79 is violated by gendered pronouns in seven intent records including a shipped intent
  • iss-338Seven registry personas are given he/she/his/her pronouns across intents/drafts (itd-92,97,98,108,126,128) and intents/shipped (itd-100), violating the they/them persona discipline (itd-79); review-enforced convention with no mechanical gate, concentrated in one 2026-08 authoring batch — a cosmetics-batch cleanup
  • itd-3abcd Loads Its Own Rules On Demand
  • itd-5Every Agent Prompt Carries Its Version, Earns Its Lock, And Survives Injection
  • itd-16`/abcd:audit` Is the Umbrella; Conversation/Edit Chain Is Its Default Application
  • itd-29Domain Experts Run A Spec Autonomously And Recover From Anything Without Touching Git
  • itd-33Multiple Agents Coordinate On Intent And Spec Work Without Duplicating Effort Or Producing Competing Artefacts
  • itd-81No Judge Ships Unmeasured, And A Clean Diff Is A Result