Skip to content

docs(architecture): define admitted executable operation packages#682

Merged
flyingrobots merged 7 commits into
mainfrom
docs/executable-warp-semantics
Jul 19, 2026
Merged

docs(architecture): define admitted executable operation packages#682
flyingrobots merged 7 commits into
mainfrom
docs/executable-warp-semantics

Conversation

@flyingrobots

@flyingrobots flyingrobots commented Jul 19, 2026

Copy link
Copy Markdown
Owner

Claim

Echo needs a new operation-oriented corridor in which admitted executable meaning is interpreted by Echo without application-supplied matcher, executor, footprint, or mutation-plan callbacks. Provider v1 remains stable callback-shaped compatibility infrastructure; this ADR does not rename it or mutate its contract.

Current callback-semantics defect

The existing provider-v1 path binds exact package, operation, Target IR, and scheduler-rule identities, but application execution meaning still enters through ambient native matcher, executor, and footprint hooks. Those bindings prove exact wiring and identities, not that the callback implements the admitted Edict meaning. A provider-v1 receipt therefore does not close the semantic-execution proof chain.

Accepted executable-operation boundary

The durable decision introduces operation-oriented nouns, including ExecutableOperationPackageV1 and subordinate EchoOperationProgramV1, while deliberately refusing to call the corridor provider v2 or to claim a universal recursive-WARP artifact.

The binding direction is explicit:

Jedit-owned Edict intent / admitted executable-operation package
→ exact Jedit canonical text-schema declaration
→ canonical semantic closure
→ exact Jedit lawpack coordinate and digest
→ exact EchoOperationProgramV1 bytes and digest

The admitted operation package owns the public operation coordinate, schemas, basis contract, budgets and footprint contract, authority requirements, result/obstruction interpretation, and runtime-recognized invocability. The program supplies subordinate executable meaning only. A program digest is not an operation coordinate, installed operation, capability, authority token, or permission to invoke. Echo starts installation and lookup from the admitted package and independently admits every invocation.

For the first ReplaceRange vertical, the selected semantic closure is a Jedit-owned, digest-locked declarative executable resource bound through a Jedit lawpack. This avoids both a handwritten semantic callback and a broad first-stage expansion of Edict into a recursive rope language.

Exact preparation and evidence laws

The review remediation freezes one canonical EchoOperationEvaluationBasisV1 encoding from existing Echo coordinates: writer head, worldline tick, optional commit-global tick, state root, commit ID, and the canonical Jedit application-basis schema/value digests.

Preparation and retained execution evidence bind:

  • exact package admission, installed operation, submission/invocation, caller-authority posture, and Echo invocation-admission evidence;
  • exact package, program, canonical input, basis, budget, footprint, scheduler-composition, and trace identities;
  • either a committable patch/output or an obstruction with canonical no-parent-patch evidence.

Over-budget evaluation obstructs. A changed basis cannot commit. Composed ticks bind the scheduler rule, ordered preparation membership, and composite patch. Only a successful terminal variant carries a committed TickPatch digest.

Verification evidence grades

The ADR distinguishes three claims:

  1. Deterministic self-validation — canonical decoding, digest recomputation, admission/basis/budget/footprint checks, scheduler composition, repeatability, and WAL reconstruction.
  2. Structurally separate verifier path — a separate path checks Core/Target/program/package correspondence, without claiming implementation independence merely because the code lives elsewhere.
  3. Independently implemented conformance evidence — finite golden/differential witnesses count only when the comparison implementation does not reuse the new encoder, lowerer, or evaluator path.

Included documentation

This PR now intentionally includes the source-grounded planning ledger. The ledger is marked as historical discovery reconciled by ADR 0023, treats the ADR as authoritative, and records the accepted four-stage campaign:

  1. minimal Edict operation prerequisites;
  2. hook-free Echo bounded-operation evaluator;
  3. Jedit ReplaceRange semantic resource and differential corpus;
  4. current Jim invocation and legacy replacement cutover.

GeneratedRules.md now distinguishes the implemented provider-v1 compatibility corridor from the required callback-free executable-operation corridor.

Explicit non-goals

This documentation decision does not add:

  • Campaign 1 implementation;
  • a Cyber Kitten runtime or syntax;
  • TextWindow or observer routing;
  • durable child lanes, child-local ticks, wormholes, holograms, or Continuum transport;
  • external-effect execution;
  • arbitrary native plugins, WASM, or a general-purpose VM;
  • broad authorization;
  • create, checkpoint, inverse, save, or causal-topology migration.

Campaign 1 implementation has not started. After this PR is explicitly authorized and merged, it begins on a fresh branch from the resulting origin/main.

Files in scope

  • docs/adr/0015-registry-provider-host-boundary.md
  • docs/adr/0023-admitted-executable-operation-packages.md
  • docs/adr/README.md
  • docs/architecture/application-contract-hosting.md
  • docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md
  • docs/topics/GeneratedRules.md
  • tests/docs/test_adr_namespace.sh

Validation

  • cargo xtask pr-preflight --base origin/main
  • bash tests/docs/test_adr_namespace.sh
  • bash tests/docs/test_retired_component_refs.sh
  • bash tests/docs/test_knowledge_model.sh
  • bash tests/docs/test_generated_rule_truth.sh
  • bash tests/hooks/test_verify_local.sh — 159 passed, 0 failed
  • Prettier, markdownlint, diff-whitespace, pre-commit, and pre-push checks
  • 99 pinned Echo/Edict/Jedit source links mechanically resolved with in-bounds line ranges

@coderabbitai

coderabbitai Bot commented Jul 19, 2026

Copy link
Copy Markdown
Contributor

Review Change Stack

📝 Walkthrough

Walkthrough

The change adds ADR 0023 for admitted executable operation packages, partially supersedes ADR 0015, updates ADR indexing and namespace validation, clarifies application-contract authority boundaries, and documents current and target Jim/Edict/Echo WARP executable semantics.

Changes

Executable semantics architecture

Layer / File(s) Summary
Authority records and compatibility boundaries
docs/adr/0015-registry-provider-host-boundary.md, docs/adr/0023-admitted-executable-operation-packages.md, docs/adr/README.md, docs/architecture/application-contract-hosting.md, tests/docs/test_adr_namespace.sh
ADR 0023 is added and indexed, ADR 0015 is marked partially superseded, namespace validation includes ADR 0023, and provider-v1 versus Echo-admitted executable operation boundaries are documented.
Package admission and execution laws
docs/adr/0023-admitted-executable-operation-packages.md
The ADR defines package/program identity binding, Echo interpretation requirements, prepared-basis validation, Jedit rope-law closure, and prohibited callback or native-intrinsic substitution.
Verification ownership and Campaign 1 sequencing
docs/adr/0023-admitted-executable-operation-packages.md
The ADR specifies evidence grades, jurisdiction across Jedit, Edict, and Echo, Campaign 1 sequencing, rejected alternatives, consequences, non-goals, and evidence anchors.
Current Jim, Edict, and Echo execution inventory
docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md
The plan inventories current operations, unsupported session capabilities, unavailable command concepts, latent runtime concepts, and non-canonical graph data.
WARP executable target and commit constraints
docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md
The plan describes bounded transactional ReplaceRange execution, canonical executable objects, receipt bindings, basis and tick-patch constraints, constitutional laws, and the recommended design campaign.

Estimated code review effort: 3 (Moderate) | ~25 minutes

🚥 Pre-merge checks | ✅ 5
✅ Passed checks (5 passed)
Check name Status Explanation
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title accurately summarizes the main change: defining admitted executable operation packages in the architecture docs.
✨ Finishing Touches
📝 Generate docstrings
  • Create stacked PR
  • Commit on current branch
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch docs/executable-warp-semantics

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9600478732

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread docs/adr/README.md
Comment thread docs/adr/0023-admitted-executable-operation-packages.md
Comment thread docs/adr/0023-admitted-executable-operation-packages.md Outdated
Comment thread docs/adr/0023-admitted-executable-operation-packages.md Outdated

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md`:
- Around line 1103-1115: Update the future receipt schema to make the committed
TickPatch digest present only for successful commits. For obstructed or rejected
evaluations, omit that field and require a canonical obstruction or outcome
digest instead, while preserving the existing result/obstruction identity
requirement.
- Around line 999-1009: Choose a single canonical executable program name
consistent with the ADR contract, replacing or reconciling WarpProgramV1
throughout the plan’s package contents and receipt identity references. If
WarpProgramV1 remains supported, explicitly document whether it is an alias,
predecessor, or distinct artifact and define its compatibility relationship to
EchoOperationProgramV1.
🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

  • Push a commit to this branch (recommended)
  • Create a new PR with the fixes

ℹ️ Review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: f9f9be04-4704-4d8f-bc80-94b4c6c4f95a

📥 Commits

Reviewing files that changed from the base of the PR and between 6615d3a and 73f73a6.

📒 Files selected for processing (6)
  • docs/adr/0015-registry-provider-host-boundary.md
  • docs/adr/0023-admitted-executable-operation-packages.md
  • docs/adr/README.md
  • docs/architecture/application-contract-hosting.md
  • docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md
  • tests/docs/test_adr_namespace.sh

Comment thread docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md Outdated
Comment thread docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md Outdated

Copy link
Copy Markdown
Owner Author

⚖️ Self-Code Review — Code-Lawyer Pass

Reviewed: 73f73a6a1a25e67c3404caba588b79b6a6d3104d against freshly fetched origin/main at 6615d3a97731a076fb4945bb6da083e82f55710d.

Pre-review gate passed: clean worktree, no staged/unstaged/untracked paths, then git fetch --prune origin.

Automated evidence is green: diff whitespace, Prettier, Markdown lint, dead references, ADR namespace, knowledge-model checks, focused docs checks, all 159 hook-regression assertions, and current GitHub Actions. I also mechanically checked all 92 pinned cross-repository source links in the ADR and plan: every commit/path exists locally and every cited line range is in bounds.

Severity summary

Severity Count
Critical 1
High 6
Medium 2
Low 1
Total 10

Summary table

ID Severity Evidence Infraction
CR-1 Critical docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md:997-1075; contrast docs/adr/0023-admitted-executable-operation-packages.md:74-138 The plan makes WarpProgramV1 contract-bearing and describes packages as installing programs, contradicting the ADR’s program-substitution boundary and selected target-relative noun.
HI-1 High docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md:1058-1075,1184-1218; contrast ADR 0023:301-316 Jurisdiction language says consequences are “admitted,” application packages “install,” and Edict rules own an “admitted application consequence.” Admission, installation, semantic authorship, and commitment are collapsed.
HI-2 High docs/adr/0023-admitted-executable-operation-packages.md:154-193 The Exact Prepared-Basis Law has no exact basis type: “frontier/tick, commit or root” is an alternative-filled list, so byte identity and equality are not implementable consistently.
HI-3 High ADR 0023:139-145,159-183; plan:1103-1120 Obstruction is specified as having no parent patch, but preparation/receipt evidence unconditionally requires a resulting/committed patch digest. The model forces either fabricated empty-patch evidence or an unbound obstruction path.
HI-4 High Plan:6-8,1235-1281; contrast ADR 0023:318-346,389-399 The plan is stale and recommends a five-witness grand design covering checkpoint, optics, save/capabilities, and fork/braid/settlement, while the accepted ADR deliberately limits the campaign and declares those areas non-goals.
HI-5 High Plan:91-96,1184-1195; contrast ADR 0023:272-299 The plan constitution requires an “independently verified” executable program even though the accepted verifier doctrine says structural separation is not implementation independence and the first vertical must not overclaim it.
HI-6 High Plan:669-718; ADR 0023:195-236,301-316 The plan proves that TypeScript and native Jedit expose materially different fact schemas, then says the fork must be resolved before authoritative .edict source. The ADR selects the lawpack closure without making this a typed refusal/precondition.
ME-1 Medium docs/adr/README.md:6-19; ADR 0023:318-346,372-374 ADR 0023 tracks campaign sequence and work scope despite the repository ADR contract explicitly prohibiting work/progress tracking and implementation diaries.
ME-2 Medium PR #682 description; branch diff The PR body still says the 1,281-line planning ledger is intentionally excluded, but it is now committed. The reviewer-facing scope and validation narrative are false.
LO-1 Low ADR 0015:7; plan:15-17,122,280,896,1128,1147 Markdown is lint-clean but inconsistent: the superseding ADR is plain text rather than a link, evidence SHAs are bare prose, and code identifiers/scalars are inconsistently formatted as prose.

Findings grouped by file

docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md

CR-1 — Program-as-contract substitution returns

The plan defines WarpProgramV1 as the executable unit and places the entry operation coordinate, input/output schemas, obstruction map, and result projection inside it. It then says an “Application package may install application WarpProgramV1.” ADR 0023 intentionally does the opposite: EchoOperationProgramV1 is target-relative subordinate executable meaning; the admitted operation package owns the public contract; Echo never installs or invokes a naked program.

This is not a naming nit. It reopens the exact substitution boundary the final ADR commit closed.

Agent-ready mitigation prompt:

Reconcile plan section 9 with ADR 0023. Treat ADR 0023 as authoritative. Replace or explicitly retire WarpProgramV1 and SemanticPackage as first-version normative nouns. Model ExecutableOperationPackageV1 as owning the public operation coordinate, invocation/output/obstruction schemas, basis and budget contracts, authority requirements, and result interpretation. Model EchoOperationProgramV1 as subordinate executable graph binding/rules/control/intrinsics only. State that Echo starts installation and invocation lookup from an admitted operation-package identity and cannot install or invoke a naked program digest. If future portable WARP nouns remain, label them non-normative hypotheses with an explicit compatibility relationship to ADR 0023.

HI-1 — Runtime jurisdictions are grammatically collapsed

“Application package may install,” “observable consequence admitted,” and “Edict-authored application rules own admitted application consequence” spend authority at the wrong crossings. Jedit owns product semantics; Edict owns authored meaning and lowering; Echo owns package/invocation admission, installation, evaluation, scheduling, and commitment. Submissions or invocations are admitted; parent-visible consequences are committed.

Agent-ready mitigation prompt:

Rewrite the plan’s jurisdiction and constitutional-law passages with exact verbs. Say that Jedit owns application semantics, Edict preserves and lowers those semantics, Echo admits packages and invocations, Echo installs admitted executable meaning, Echo evaluates/schedules, and Echo commits parent-visible consequences. Replace “consequence admitted” with “consequence committed.” Replace “application package may install” with Echo-owned installation eligibility through admitted package evidence. Do not let an Edict artifact or application package own Echo runtime authority.

HI-4 — The committed plan is stale and reintroduces the grand-unification campaign

The header says implementation is “halted pending architectural agreement,” and section 13 still asks for the architecture document that this PR now supplies. Worse, it requires five witnesses spanning checkpoint, optics, capabilities, and causal topology. ADR 0023 explicitly keeps those out of the first campaign.

Agent-ready mitigation prompt:

Update the plan as a historical discovery record resolved by ADR 0023. Add a prominent status/link identifying ADR 0023 as the accepted decision and state that Campaign 1 implementation has not started. Replace the obsolete five-witness design campaign with the accepted four-stage sequence: minimal Edict prerequisites; hook-free Echo evaluator with a tiny generic program; Jedit ReplaceRange lawpack artifact and differential corpus; Jim invocation/cutover. Explicitly preserve checkpoint, TextWindow, save/capabilities, Kitten runtime, durable child lanes, wormholes, and causal-topology migration as non-goals. Remove any instruction to design the grand unification before the first seam.

HI-5 — Verifier independence is overclaimed

The plan elevates “independently verified executable WARP program” into constitutional law. The ADR correctly states that a separate crate/path is not an independent implementation and reserves independent conformance claims for genuinely separate finite witnesses.

Agent-ready mitigation prompt:

Align every verifier claim in the plan with ADR 0023’s three evidence grades. Use “deterministic self-validation” for same-implementation checks, “structurally separate verifier path” for separate target checking, and “independently implemented conformance evidence” only for comparisons sharing no encoder/lowerer/evaluator implementation. Remove unconditional “independently verified” wording from the first-version constitution, or mark it as a future stronger goal that is not claimed by Campaign 1.

HI-6 — The proven Jim schema fork has no ADR refusal gate

The plan documents ten TypeScript facts versus eight native facts, with materially different meanings—not mere encodings—and correctly warns that choosing one accidentally would canonize an unresolved semantic fork. ADR 0023 requires an “exact fact/codec/identity law” but does not say generation/admission must stop until Jedit selects the authoritative propositions and migration posture.

Agent-ready mitigation prompt:

Add a Jedit-owned schema prerequisite to ADR 0023’s rope-law closure. Before the real ReplaceRange resource can be generated or admitted, require one canonical fact/edge/attachment schema and identity/codec law, an explicit decision for the TypeScript-only maintenance/anchor facts, and a migration or compatibility posture for the native JSON model. Require typed refusal if the selected lawpack, program, schemas, and Jedit authority declaration do not close exactly. Do not let Edict or Echo choose between the existing models.

LO-1 — Inline-code and evidence-link style is inconsistent

The document passes Markdown lint, but many code identifiers and scalar types are plain prose, and the source-head SHAs are not code-formatted or linked. This makes a source-grounded architecture report less scannable than its own evidence citations.

Agent-ready mitigation prompt:

Perform a Markdown-only style pass over the plan. Wrap code identifiers, operation coordinates, field names, scalar types, and ABI/domain names in inline code consistently. Turn source-head SHAs into pinned repository links or inline-code coordinates. Preserve wording and semantic content; run Prettier, markdownlint, and dead-reference validation afterward.

docs/adr/0023-admitted-executable-operation-packages.md

HI-2 — “Exact basis” is not represented exactly

The ADR binds “worldline, writer head, frontier/tick, commit or root, and application basis.” The slashes and “or” leave multiple lawful implementations. The equality pseudocode compares an undefined conceptual noun rather than canonical bytes or a versioned value.

Agent-ready mitigation prompt:

Define the first-version canonical evaluation-basis value used by PreparedEchoOperationV1. Inventory the existing Echo/Jedit basis coordinates, then specify a versioned required tuple or digest with no alternatives: exact worldline/runtime frontier coordinate, writer-head coordinate where applicable, and the application-declared Jedit basis relationship. Define canonical encoding, identity domain, equality, and typed mismatch evidence carrying evaluated and current identities. State whether every field is mandatory or selected by an exact admitted target profile; do not leave “frontier/tick, commit or root” as prose alternatives.

HI-3 — Success and obstruction need an explicit sum type

The ADR promises one patch or one obstruction and no partial parent mutation, but its evidence list requires a patch digest for every preparation and precommit check. The plan repeats the problem by listing a committed patch digest alongside “result or obstruction.”

Agent-ready mitigation prompt:

Replace the implicit prepared result with an explicit closed outcome enum. A committable outcome binds patch digest and typed output and may enter exact-basis commit. An obstructed outcome binds the typed obstruction, consumed budget, actual footprint/read support, trace identity, and canonical evidence that no parent patch exists; it never runs the committed-patch equality check. Update receipt and WAL obligations accordingly. Add a negative witness proving obstruction cannot fabricate, retain, or partially apply a parent patch.

ME-1 — ADR 0023 violates the repository’s ADR contract

The ADR index says ADRs do not track work/progress and must not become implementation diaries. “Sequencing Consequence” and repeated Campaign 1 scope prose are roadmap state. Now that a plan file is deliberately committed, that state has a proper home.

Agent-ready mitigation prompt:

Reduce ADR 0023 to durable decisions, alternatives, jurisdiction, invariant laws, consequences, and non-goals. Move the numbered campaign sequence, Campaign 1 implementation boundary, and current work status into the plan and PR/issue ledger. Keep only the architectural dependency rationale necessary to explain the decision. Preserve links between the ADR and plan, and rerun ADR namespace/Markdown/dead-reference checks.

PR #682 metadata

ME-2 — The PR description is factually stale

The description says: “The larger local planning ledger is intentionally not part of this PR.” The ledger now contributes 1,281 lines, and the PR also modifies the ADR namespace shell test.

Agent-ready mitigation prompt:

Update PR #682’s title/body after the branch fixes. Remove the claim that the planning ledger is excluded. Enumerate all six changed files, explain why the source-grounded plan is included, state that the shell change is a documentation namespace guard, and record the focused namespace and 159 hook-regression checks. Keep the statement that Campaign 1 implementation has not started.


@codex Please independently confirm or challenge each finding against the current head 73f73a6a1a25e67c3404caba588b79b6a6d3104d. In particular, verify the program-substitution conflict, exact-basis under-specification, success/obstruction evidence sum, verifier-grade claims, and the Jedit schema refusal gate before this PR is treated as merge-ready.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 73f73a6a1a

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md Outdated
Comment thread docs/adr/0023-admitted-executable-operation-packages.md Outdated
Comment thread docs/adr/0023-admitted-executable-operation-packages.md Outdated
Comment thread docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md Outdated
Comment thread docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md Outdated
Comment thread docs/plans/2026-07-18-jim-edict-echo-executable-warp-semantics.md Outdated
Comment thread docs/adr/0023-admitted-executable-operation-packages.md
Comment thread docs/adr/0023-admitted-executable-operation-packages.md Outdated
Comment thread docs/adr/0023-admitted-executable-operation-packages.md Outdated
Comment thread docs/adr/0023-admitted-executable-operation-packages.md Outdated
@flyingrobots
flyingrobots merged commit 3e1e47a into main Jul 19, 2026
39 checks passed
@flyingrobots
flyingrobots deleted the docs/executable-warp-semantics branch July 19, 2026 07:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant