chore(release): version packages

[openspec-release-bot]

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

@fission-ai/openspec@1.9.0

Minor Changes

• #1622 59c16a4 Thanks @clay-good! - ### New Features

  • Command Code command adapter — Command Code is now a first-class, adapter-backed tool. openspec init generates OpenSpec commands under .commandcode/commands/opsx-<id>.md (invoked as /opsx-<id>) alongside the skills under .commandcode/skills/, matching Command Code's documented custom-slash-command surface.

• #1613 42d7f67 Thanks @Angelthebestone! - ### New Features

  • Command Code support — openspec init now supports Command Code as an adapterless skills-only tool. It installs the OpenSpec skills under .commandcode/skills/ and invokes them as /openspec-* commands, matching Command Code's native skill surface.

• #1604 83be9d1 Thanks @clay-good! - Add openspec validate --archived: an opt-in check that every change under changes/archive/ has all of its tasks.md checkboxes ticked, exiting non-zero if any are unchecked. This surfaces changes that were archived with unfinished work — which the normal validate flow never catches, because it only looks at active changes — and is meant for a pre-commit or CI hook (#205). It is a standalone scope: it does not alter any existing validate invocation and does not re-validate already-applied spec deltas.

Patch Changes

• #1530 bf5099e Thanks @clay-good! - Apply workflow now tells agents to surface unexpected scope instead of hiding it. When a task needs work beyond what the spec describes, the /opsx:apply skill and command guidance direct the agent to pause and report the added scope rather than silently narrowing, deferring, or simplifying away specified behavior, and to mark a task complete only when its specified behavior is fully implemented. Fixes #1529.

• #1603 9ae75c8 Thanks @clay-good! - openspec archive no longer writes terminal escape codes to a redirected or captured stdout. Its confirmation prompts and the no-argument change picker drew their live UI with ANSI cursor-move sequences even when stdout was not a terminal — noise in a redirected log, and in some non-interactive hosts an unbounded render loop that could grow the captured output until the disk filled. When stdout (or stdin) is not a terminal, archive now reads the confirmations as plain text, and a no-argument run asks you to pass a change name up front instead of drawing a menu. Piped answers (printf 'y\n' | openspec archive …) and --yes behave as before, and interactive terminals are unchanged. Fixes #1526.

• #1528 9425897 Thanks @Marzx13! - Canonicalize rebuilt specs to end with exactly one final LF. Previously a spec whose ## Requirements section was last was rebuilt with a trailing blank line (\n\n), which failed Markdown whitespace checks after sync or archive. Internal spacing and content after the Requirements section are unchanged.

• #1640 610b78f Thanks @clay-good! - Preserve the blank lines around a spec's ## Requirements heading when syncing a delta. openspec archive rebuilt openspec/specs/<capability>/spec.md by joining its slices with a bare newline, so the blank lines that surround the heading were dropped and the resulting file failed Markdown whitespace checks. The rebuild now keeps that spacing intact. Fixes #1625. Thanks @jwang513! (#1637)

• #1640 610b78f Thanks @clay-good! - openspec validate --all and openspec list --json no longer silently pass when run outside an OpenSpec project. From a directory with no root they used to resolve the current directory as an implicit root, exit 0, and report empty results — a false pass for CI and agents. Bulk validation (--all, --changes, --specs) and list now require an existing root (the openspec/project.md fallback for legacy projects is kept), while direct validation and other intentional implicit-root workflows are unchanged. Thanks @clay-good! (#1612)

• #1640 610b78f Thanks @clay-good! - Label the update workflow in the openspec config workflow picker. The checklist had friendly labels for 11 of the 12 workflows but was missing update, so that row — one of the six core workflows every user sees — fell back to its raw id with a placeholder description. The update-change template's stale "expanded-profile" wording is also reworded to "optional". Fixes #1627. Thanks @clay-good! (#1632)

• #1640 610b78f Thanks @clay-good! - openspec schema fork now preserves the source schema's YAML formatting. Renaming a forked schema.yaml round-tripped through a parse/re-serialize step that dropped comments, could rewrite block-scalar style (a literal | folded to >), and reordered keys, so the fork no longer matched its source. The rename now edits the document in place via the YAML Document API, leaving comments, scalar style, and key order untouched. Thanks @clay-good! (#1607)

• #1640 610b78f Thanks @clay-good! - openspec schemas now resolves through the canonical OpenSpec root-selection precedence instead of always reading from the current directory. It accepts --store <id>, rejects --store-path like the other store-aware commands, and returns the shared machine-readable diagnostics on JSON failures, while preserving the existing human output and bare JSON array on success. Thanks @Patodo! (#1616)

• #1640 610b78f Thanks @clay-good! - openspec validate now warns on ambiguous task numbering in spec-driven changes: a task ID duplicated at full depth (including across resolved task files), or a task whose leading number disagrees with its enclosing ## N. group. Numeric-looking text outside numbered groups is ignored, and custom schemas are unchanged until they opt in. The checks run across direct, bulk, and deprecated change validation. Closes #1520. Thanks @alectimison-maker! (#1523)

• #1522 07dea6e Thanks @clay-good! - ### Bug Fixes

  • Don't let a legacy Codex upgrade hijack the vendor-neutral agents target — openspec update no longer overwrites an existing .agents skills tree (and its ownership marker) when Codex is detected only from leftover global ~/.codex/prompts. Because Codex and the vendor-neutral agents target share .agents/skills, a project that used the agents target could have its generic skills silently rewritten with Codex-specific syntax and its target flipped to Codex on the next update --force. The legacy-upgrade path now respects the established owner of a shared skills directory, matching the one-writer rule openspec init already applies. When an upgrade is skipped this way, that tool's repo-local legacy files (e.g. .codex/prompts/openspec-*.md) are also preserved rather than cleaned up, since no replacement was written to take their place. A genuine first-time Codex upgrade (no .agents tree yet) is unaffected.

• #1521 c751b3d Thanks @clay-good! - ### Bug Fixes

  • Stop silently dropping unlabeled scenarios on archive — openspec validate and openspec archive now recognize every level-4 (#### followed by whitespace) child of a requirement as a scenario, matching how the spec is counted elsewhere. Before, the scenario-loss guard only recognized headers written exactly as #### Scenario:, so a MODIFIED requirement that dropped a differently-labeled child (for example #### Edge case) passed validation and was then permanently deleted by archive with no warning. Both paths now agree, so the loss is caught at authoring time. Scenario names are normalized when comparing (an optional Scenario: prefix and a CommonMark closing # run are ignored), so simply relabeling a scenario is not mistaken for dropping one.

• #1610 17581c1 Thanks @clay-good! - ### Bug Fixes

  • openspec init now suggests an IDE restart only when an IDE-resident tool such as Cursor, GitHub Copilot, Continue, or Cline was configured. CLI tools like Claude Code, Codex, and Gemini CLI no longer show the hint, since their commands work as soon as the files exist.

• #1609 804427b Thanks @clay-good! - Suppress the first-run telemetry disclosure notice when --json is used. On a
  first-ever run the notice was written to stdout and could break --json
  consumers; it is now deferred to the first later non-JSON run, keeping --json
  output valid while still guaranteeing the disclosure.