Skip to content

chore: version packages - #173

Open
github-actions[bot] wants to merge 1 commit into
mainfrom
changeset-release/main
Open

github-actions[bot] wants to merge 1 commit into
mainfrom
changeset-release/main

Conversation

@github-actions

@github-actions github-actions Bot commented Nov 17, 2026 •

Copy link
Copy Markdown
Contributor

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

@stainless-code/verdict@0.8.0

Minor Changes

  • ecc6b89: drain is a hidden CLI verb that refreshes threads, plans the review threads waiting on a verdict reply, and pins the plan with a digest. drain --agent <name> launches one agent through the same launcher run uses and validates its drain.replies.json against the plan. No host writes or reply posting ship in this slice.

  • 6923d0e: drain --post answers the waiting review threads from a validated drain.replies.json. GitHub and GitLab post one marked reply per draft and append replies.json under a lock after each write; withdraw also resolves its thread. Secret-shaped bodies refuse before any write, stale threads are skipped, and a failed write records failed[] and refuses drain-incomplete so the same command can finish the remaining replies.

  • 9c68ea2: run --agent now supports fourteen verified official agent CLIs, including agy, amp, copilot, crush, droid, gemini, goose, kiro, pi, and qwen, with prompt flags placed immediately before the directive for agents that take it as a flag value. Operators can define more agents under the new agents key of their own settings file; the shape is prefs-only, closed, and refuses name collisions and invalid binaries as bad-settings.

  • bc6f739: A bad --context path now refuses context-invalid with a path-specific hint instead of fetch-failed.

    Breaking changes

    A bad --context input — missing, symlinked, too deep or too large, or the wrong type — now refuses context-invalid instead of fetch-failed. Scripts that matched fetch-failed for these cases must match context-invalid. Bundle-write failures still refuse fetch-failed. The refusal names the path as typed and, for a relative path, its absolute resolution, and its hint tells you to check each --context path.

  • 98d8bec: run --agent now offers 32 more cited headless agents, including Cline and Junie after CLI-reference checks. Rows whose binary shares a name with another tool run a harmless identity probe before launch; a non-zero exit, a five-second timeout, or a missing match refuses the new agent-mismatch code before any agent starts, and launcher.json records probe: {args, ok} for a passing check. Operator agents can set probe_args and probe_match under settings.json agents, with probe_match required when probe_args is present.

  • 770b32f: drain --post --dry-run runs every post gate and live read, then writes drain.dryrun.json with the exact marked replies and typed skips a real post would produce. It makes zero host writes and leaves replies.json, drain.json, and threads.json byte-identical; --agent --post --dry-run plans, drafts, then previews. The flag requires --post and secret-shaped bodies still refuse.

  • 5712cec: Breaking changes

    The installed skill is now a small Bootstrap skill: it only says when to use Verdict and how to reach the binary. The binary serves the full Procedure, so the instructions always match the version that checks your work.

    Reinstall the skill on every machine. For harness installs, for each installed agent, run verdict skill status --agent <x> and delete the old verdict skill directory at the path printed after ->; then run verdict skill install --agent <x>. verdict skill install --detect is fine instead of --agent. verdict skill uninstall removes only the new Bootstrap skill files, so it would leave the old Procedure files beside the stub.

    For registry installs, delete the old verdict skill directory, then re-run npx skills add stainless-code/verdict.

    Copies went stale between releases: the Procedure lived inside the installed skill, so an older copy could hand an agent instructions that disagreed with the binary grading its work. Keeping it in the binary removes that skew.

  • 66342c5: Every fetched bundle now carries procedure.md, the review Procedure assembled by the binary from its embedded payload. Drain plans carry drain.procedure.md for reply drafting. The run, stage, and drain directives name those files as authoritative, so an installed skill that is older can no longer displace them. The fetch next: hints and --show-prompts output point at the bundle Procedure, and provenance receipts record its sha256 as a bundle input. PromptsHash now reads the embedded payload, so an installed user gets the real prompt-set hash without a procedure/ checkout.

  • 85b2e97: MCP hosts can now load the Procedure straight from the binary. prompts/list and prompts/get serve five task prompts (address-reviews, fix-ci, incremental-review, review, validate), and resources/list and resources/read serve every payload file at verdict://skill/<path>, byte-identical to verdict skill show. Prompts and resources are read-only and available in every server mode, including --read-only and --operator. initialize advertises both capabilities and points hosts at the review prompt or verdict://skill/PROCEDURE.md.

  • 31efb1b: Breaking changes

    Published JSON schemas moved from skills/verdict/schemas/ to /schemas/; every $id changed accordingly.

  • 9de3880: Verdict now writes an append-only per-bundle invocation journal under <state>/verdict/journal/<repo-key>/<pr>.jsonl ($XDG_STATE_HOME, falling back to ~/.local/state on Unix or %LOCALAPPDATA% on Windows) for every bundle-writing CLI and MCP call, and adds the read-only verdict journal <owner/repo> <pr|local-name> [--json] verb (plus its MCP tool, Go facade method, TypeScript client method, shell completion, and schemas/journal.schema.json). Each call records paired start and end lines carrying the canonical repo root, redacted argv, content-digest bundle.before/bundle.after tamper-evidence hashes, exact wrote/removed paths with SHA-256 digests, exit code, refusal code, warnings count, and next hint, bounded at 1 MiB with atomic whole-line truncation and non-fatal append warnings. The folded verdict journal summary pairs each call under a unique step key, reports the size-cap notice under truncation, and publishes bundle.unobserved with no file diff when a content sweep fails instead of claiming every file changed. Because the journal lives outside the repository and .verdict/, dry runs, refusals, and bundle receipts stay completely untouched.

  • 938bc6d: review.json is now a findings-only handoff: each finding carries a path, strict integer line, severity, claim, quoted code, and finder confidence; optional spec checks and receipts reserve the evidence S4 verifies. Verdict renders the body and derives approval from the stored findings, with a documented default threshold and the existing block-clause policy. The document keeps "schema_version": "1" — this is a breaking change inside that version, shipped with no shim.

    Breaking changes

    review.json must carry findings[] and may no longer carry approved, body, comments[], sections[] or a top-level confidence; the document is closed, so any of those keys refuses as an unknown field and there is no repair hint for the old shape. To rewrite a stored v1 file, move each comments[].body into findings[].claim, add quote, severity and confidence to every finding, convert line to an integer, move the old body prose to summary (at most 2000 characters), drop approved, sections[] and top-level confidence, and keep the claim arrays (resolved, triage, coverage, chunks, suppressed, risk, security_summary) unchanged.

    One problem code that policed the old contract is gone: basis-replaced (the agent writes no
    body, so there is no line for the CLI to repair). The closed document refuses the removed keys
    as unknown-field, and the review prompt drops in-bounds with the rest of the body grammar.

    The dry run's data.check now reports findings and anchors_moved_to_body in place of the
    comments and sections counters that described the old body grammar.

    Approval is no longer an agent claim. With no repository policy, any open, visible finding whose quote standing is not unverified at important or above withholds approval, as does any violated spec_checks[] entry; block_clauses and report --check refine that default when configured. record decides once and stores the decision in record.json as approval, beside the review_sha256 it decided on; post publishes that stored decision, renders its body from the same decision, and refuses record-stale when the review, diff, or head moved after the record, rather than deciding again over a review nobody recorded.

    Do not release between this slice and S4. S1 stores spec_checks[] and receipts[] but does not verify quotes or require receipt coverage; S4 adds the insufficient-review refusal and closes the gap where a findings-free review would otherwise derive approval. The release hold is tracked in Linear.

  • 1dfcc29: backtest fixtures grades converted fixture recordings against hand-known defects in a new fixture corpus. It publishes recall, bonus-union, false-green, hallucination (with explicit unverifiable counts) and review-insufficient metrics, reproduces seven hand scores as calibration, and writes one report under .verdict/backtest/fixtures/report.json. The mode never launches an agent or calls a host.

  • 4ab8691: The review handoff is now lean and layered. <bundle>/procedure.md carries a pushed core — role, two method lines, the findings-only contract, and a capability map — capped at 3 KB by an archtest. Sections for chunked diffs, prior threads, knowledge-base rules, lenses, and handed context are appended only when the bundle carries that fact. A sliced review checklist and the remaining reference material are written to <bundle>/refs/ and pulled on demand; the full operator workflow and body-format rules stay with verdict skill show. --show-prompts prints the assembled handoff and one line naming the refs, never their bodies. Drain and the MCP task prompts use the same core where it applies.

    The directive now says to write review.json first with empty findings and update it as the review proceeds; after a clean exit only (no crash, timeout or non-zero exit), run salvages the last fenced json block on stdout that decodes to an object with a findings array, adding schema_version "1" if missing, records salvaged: true in launcher.json, and warns. It replaces only an absent artifact or the exact empty placeholder; a salvaged or written review that fails the shared checker is refused bad-review-input with its problems, which feed the next round under --rounds, while only stdout with no such block keeps no-review-artifact.

    A finished review with no findings must carry a non-empty summary; the empty placeholder alone is never a clean review, and run checks every artifact with the same checker record uses before reporting success.

    prompts_hash and the provenance receipt now cover every file that feeds the assembled handoff and its refs, so an edit to any of those bytes is visible in the bundle's recorded prompt set.

  • 8b159e9: fetch and fetch-local now accept a repeatable --spec <file> flag to hand
    specification documents into the bundle separately from plain --context
    reference material.

    • Validation matches --context (size, type, symlink, and depth caps).
    • Spec sources are stored in the same context/ directory and fenced
      context.txt, while context.json records "role": "spec" or
      "role": "context" per entry in place and toc.txt's ## context section
      marks spec rows with [spec].
    • Combined spec sources are capped at 150 enumerated checklist items; exceeding
      the cap refuses spec-too-large with no bundle files written.
  • 5e1ab56: Verdict now checks the evidence a review states before it decides anything: each finding's quote is matched against the pinned diff at its path:line, a review whose findings are all below the approval tier must receipt every non-exempt changed hunk, and --spec sources get a per-item checklist. The verdict is stored as an evidence receipt in record.json and read by every verb: record --dry-run names each gap, run repairs them across its rounds and refuses review-insufficient when they survive, report --check and gate fail an always-on evidence dimension, and post publishes the findings with an insufficient lead and never approves.

    Breaking changes

    A review that states no evidence is now insufficient rather than approved. review.json must quote the exact line at each finding's path:line (start_line plus one quoted line per range line for a multi-line quote), and a review with no gating finding must carry receipts[] — one {"hunk":"<the hunks.tsv row>","quote":"<a line that hunk carries>"} per non-exempt hunk; verdict record --dry-run names every gap and now ends a clean review with a warning instead of silence. run --rounds defaults to 2 instead of 1, so an invalid review is re-prompted once unless you pass --rounds 1; run also refuses review-insufficient where it used to hand off, and a plain run envelope now carries data.repair on that default. report --check and gate always grade a seventh dimension, evidence, so a bundle that never recorded a receipt for its head reads unknown and fails instead of passing on the other six. fetch --spec, and the defaults.spec_coverage_min default of 80 it feeds, become part of the review contract: an enumerated spec item with no check, or with test-output evidence (logs, goldens, snapshots, testdata; test source counts as code), withholds approval. There is no compatibility shim for any of this — v0.7 is not yet at 1.0, and the previous behaviour approved reviews nothing had checked.

  • abb3389: ready --filter now gates the review on the closed trigger predicates in the trusted base configuration: matching PRs run and record the decision, non-matching PRs refuse filtered-out before any host call, and unusable policy or facts refuse trigger-failed.

    Fetch now requests and records PR labels in PR.json for every host, so the labels trigger predicate can read them from the bundle.

Patch Changes

  • 8dbe092: The --spec checklist cap is now 150 items. The 60-item ceiling refused real specification documents mid-PR; fetch and fetch-local still refuse spec-too-large before writing any bundle file once the combined --spec sources enumerate more than 150 items, and the refusal message and hint name the new limit.
  • 3232b0f: report --check and gate now fail on a violated spec check. The new always-on spec dimension reads the spec checks the stored record graded: a violated one fails the gate with violated spec check: <item> and a tally of every violated item, while met, n/a, and a review with no spec checks pass. Before this, a bundle whose only defect was a violated spec check passed the gate while the derived approval withheld with violated spec check, so the gate certified a review the decision refused.
  • 154107e: The run directive names the handed --context spec.
  • f44be4c: The MCP server now serves build, ci, and grade as task prompts, one per task prompt the
    Procedure's mode router names, so prompts/list offers eight instead of five. Each body is the
    same assembly the other task prompts use — PROCEDURE.md, prompts/system.md, and that task's
    own prompt — and an unknown name still refuses -32602 and lists the valid tasks.

@coderabbitai

coderabbitai Bot commented Nov 17, 2026 •

Copy link
Copy Markdown

Important

Review skipped

Bot user detected.

Ask an organization administrator to verify that this bot has an active assigned CodeRabbit review seat. See the seat-assignment guide. After the seat is confirmed, request a review again.

Run configuration

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

@github-actions
github-actions Bot force-pushed the changeset-release/main branch 21 times, most recently from fd84953 to 92192c0 Compare November 19, 2026 22:12
@github-actions
github-actions Bot force-pushed the changeset-release/main branch from 92192c0 to 0463198 Compare November 19, 2026 23:48

This branch has not been deployed

No deployments
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.

0 participants