Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

QVAC Agent Stack Sync · tetherto | DecimalAI

tetherto/qv-agent-stack-sync

SkillScore52Maintained

Plan and prepare the QVAC agent-stack release cascade across @qvac/sdk, @qvac/cli, @qvac/ai-sdk-provider, @qvac/opencode-plugin, and @qvac/openclaw-plugin. Use for /qv-agent-stack-sync --plan, --prepare-cascade, or --promote; syncing CLI to a newer SDK; or checking OpenCode/OpenClaw/provider compatibility.

DevOpsby tetherto

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

DevOps & SRE pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/tetherto-qv-agent-stack-sync/SKILL.md

decimalai skills pull tetherto-qv-agent-stack-sync --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/tetherto-qv-agent-stack-sync/SKILL.md https://app.decimal.ai/s/tetherto-qv-agent-stack-sync/SKILL.md

Raw file: /s/tetherto-qv-agent-stack-sync/SKILL.md ↗ · machine-readable summary at /s/tetherto-qv-agent-stack-sync.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 7d ago

—+83% tokens0% turns52% of whole cases pass with skill (12/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-08	✗→✓	▲ Improved	47%	0%
case-12	✗→✓	▲ Improved	51%	0%
case-18	✗→✓	▲ Improved	54%	0%
case-19	✗→✓	▲ Improved	80%	0%
case-20	✗→✓	▲ Improved	61%	0%

Showing 5 of 23 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

tetherto / qvac

.cursor/skills/qv-agent-stack-sync/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/10/2026Version v2Bundled 1 file

Skill contents

references/

SKILL.md

QVAC Agent Stack Sync

Orchestrate the published dependency cascade:

text
@qvac/sdk
  -> @qvac/cli
    -> @qvac/ai-sdk-provider
      -> @qvac/opencode-plugin
      -> @qvac/openclaw-plugin

Goal: know which packages need a release, prepare draft release + backmerge PRs when asked, never auto-publish. SDK releases still run qv-sdk-lockstep-sync for bare-sdk + sdk-python.

Modes

| Mode | Mutates? | When | | --- | --- | --- | | --plan (default) | No | Always run first. Cascade matrix + version recommendations. | | --prepare-cascade | Yes (confirm first) | Open draft release PRs and draft backmerge PRs for needs_release packages. | | --promote <slug> | Yes (confirm first) | Mark a draft release PR ready after lower-layer npm is live; re-verify. Does not publish. |

Invoke: /qv-agent-stack-sync or /qv-agent-stack-sync --plan.

References

.cursor/rules/sdk/sdk-pod-packages.mdc — pod scope + agent-stack order
.cursor/skills/qv-sdk-changelog/SKILL.md
.cursor/skills/qv-sdk-pr-create/SKILL.md
.cursor/skills/qv-sdk-backmerge/SKILL.md
.cursor/skills/qv-sdk-lockstep-sync/SKILL.md (sdk releases only)
docs/architecture/AGENT-INTEGRATIONS.md
packages/cli/test/AGENT_STACK_E2E.md
Planner: .cursor/skills/_lib/sdk/agent-stack-plan.mjs
Path map: scripts/sdk/package-paths.cjs
Cascade details: references/prepare-cascade.md

Version policy (pre-major 0.x)

Recommend with rationale — never silent-bump. Confirm before --prepare-cascade.

0.x carets do not cross minors (^0.8.0 ≠ 0.9.0).
Minor — outstanding [bc], [api], or feat: commits.
Patch — dep-range alignment, or other non-breaking outstanding commits (e.g. fix:).

Blockers the plan must surface

AI SDK major mismatch — e.g. provider on ai@7 while OpenCode still depends on ai@6 / provider ^0.2.x. Hold that plugin release until code + deps align.
Missing git tags vs npm latest — stop and ask; do not guess changelog base.
Lower layer not on npm yet — drafts may declare future ranges; --promote / publish wait on npm view.

Workflow

0. Shared preflight

Prefer a clean worktree / dedicated worktree for cascade prep.
git fetch org remote (upstream / tetherto/qvac) tags + main.
Identify org remote vs fork remote (same policy as qv-sdk-pr-create).

1. `--plan` (mandatory first step)

bash
node .cursor/skills/_lib/sdk/agent-stack-plan.mjs
node .cursor/skills/_lib/sdk/agent-stack-plan.mjs --json

Show the markdown table to the user. For each package report:

local version, npm latest, last tag
needs_release / blocked
suggested next version + rationale
dep ranges vs target lower version
outstanding non-noise commits since tag

Do not create branches or PRs in this mode.

2. `--prepare-cascade`

Requires explicit user confirmation of the plan (versions + which packages).

Then for each needs_release package in dependency order, follow references/prepare-cascade.md:

Create org release-<slug>-<version> from main if missing.
Prep version bump + dep ranges + changelog (qv-sdk-changelog --package=<slug>) + NOTICE.
Open draft release PR → release-<slug>-<version>.
Open draft backmerge PR → main ([skiplog], cherry-pick -x) in the same session.
Skip packages marked blocked; report them clearly.
Do not merge. Do not trigger publish (Dima / human).

SDK releases still chain qv-sdk-lockstep-sync + docs Step 8 from qv-sdk-changelog.

3. `--promote <slug>`

Confirm lower dependencies are on npm (npm view).
Fresh-install verification (below).
Mark the draft release PR ready for review (and keep backmerge draft until release merges, or ready it alongside — prefer ready both when release is ready to merge).
Remind: human merges; human triggers publish; then promote the next upper package.

File updates (when preparing releases)

CLI

packages/cli/package.json — version + dependencies["@qvac/sdk"]
changelog via qv-sdk-changelog --package=cli
NOTICE via qv-notice-generate cli
README committed SDK range if documented

AI SDK Provider

packages/ai-sdk-provider/package.json — version + extend peerDependencies["@qvac/cli"]
changelog + NOTICE via SDK pod tools

OpenCode / OpenClaw plugins

plugins/opencode/package.json or plugins/openclaw/package.json (+ openclaw.plugin.json version when present)
bump @qvac/cli and @qvac/ai-sdk-provider deps to the planned minors
OpenCode may also need ai / @ai-sdk/openai-compatible majors to match the provider — if not aligned, plan marks blocked
changelog: qv-sdk-changelog --package=opencode-plugin or --package=openclaw-plugin (path map resolves plugins/*)
NOTICE + README compatibility lines

Verification

Per package (from package dir)

CLI: lint, build, test:unit, test:e2e, node scripts/check-publish-ready.cjs
Provider: lint, build, test:unit (optional QVAC_INTEGRATION_TEST=1 npm run test:integration)
OpenCode / OpenClaw: lint, build, test:unit (optional integration where available)

Fresh install (after lower npm publish)

bash
tmp=$(mktemp -d)
cd "$tmp"
npm init -y
npm install --no-fund --no-audit @qvac/opencode-plugin@latest
npm ls @qvac/opencode-plugin @qvac/ai-sdk-provider @qvac/cli @qvac/sdk

Repeat with @qvac/openclaw-plugin@latest when promoting OpenClaw.

Completion report

Always end with:

plan table (or link to --json output)
packages released / drafted / blocked / skipped
confirmed versions and dep ranges before → after
draft release + backmerge PR URLs
tests run
reminder that publish is human-gated

Quality checklist

] --plan run before any mutation
] User confirmed versions for --prepare-cascade
] OpenClaw included in the cascade
] Draft release and draft backmerge opened together per package
] Blocked plugins (e.g. AI SDK mismatch) not force-released
] No publish / workflow_dispatch for npm
] Org-branch heads preferred (tetherto/qvac)

Want to see if tetherto/qv-agent-stack-sync would activate on your phrasing? Try it before you install — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/22/2026, 17:20 UTC · 23 cases

Measured on 16 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

+22%

52% vs 30%

Agent turns

1.0 vs 1.0

Tokens

+83%

2,258.56 vs 1,231.63

fail → pass · 5pass → pass · 7fail → fail · 11

Per-case results (23)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 We have pending commits across our QVAC packages and need an overview of what needs to be released. Please analyze the repository status against published npm versions across the SDK, CLI, AI provider, and plugins. Provide a clear breakdown showing recommended version bumps, dependency alignments, and any blocking issues, but do not create any branches or open pull requests yet.	fail→fail	3,763	6,374	+69%	1	1	0%	405	2,234	+452%	0	0	—
case-02 I've approved the proposed updates for the packages requiring a release in the stack cascade. Go ahead and generate the draft release pull requests along with matching draft backmerge pull requests for all packages ready to bump. Make sure to run the required local validations, report the version changes, and give me a summary with all PR links.	fail→fail	5,984	4,923	-18%	1	1	0%	427	2,104	+393%	0	0	—
case-03 The underlying CLI release is now live on npm, so we need to advance the draft PR for the OpenCode plugin. Please verify that the registry reflects the updated dependency, perform a fresh package installation verification in a clean temporary folder, mark the release draft as ready for review, and list the final test results.	fail→fail	11,689	5,775	-51%	1	1	0%	1,426	2,236	+57%	0	0	—
case-04 We are ready to publish @qvac/cli version 0.9.0 to npm registry. Please run npm publish --access public and push the git tags directly to GitHub.	fail→fail	3,951	7,050	+78%	1	1	0%	588	2,185	+272%	0	0	—
case-05 We have several git commits in our project repository that don't conform to conventional commit specs. Please rebase and rewrite the commit messages for the last 5 commits on main.	fail→fail	7,772	6,238	-20%	1	1	0%	1,427	2,138	+50%	0	0	—
case-06 Our frontend React component in apps/web fails to compile due to type errors in useState hooks. Please fix the TypeScript errors in components/Header.tsx.	fail→fail	12,444	6,165	-50%	1	1	0%	2,468	2,133	-14%	0	0	—
case-07 We need a machine-readable JSON summary of our package stack release status across @qvac packages. How should the planner script be invoked?	fail→fail	15,904	2,640	-83%	1	1	0%	938	2,300	+145%	0	0	—
case-08 We have releases pending across the agent stack. What order must draft releases be prepped starting from the core SDK down to downstream agent plugins?	fail→pass	12,978	9,294	-28%	1	1	0%	1,902	2,799	+47%	0	0	—
case-09 The AI SDK provider package is updated to support AI SDK version 7, but our OpenCode plugin repository still references AI SDK version 6 and provider ^0.2.x. How does the release planner mark the OpenCode plugin status?	pass→pass	9,007	4,469	-50%	1	1	0%	1,317	2,575	+96%	0	0	—
case-10 We have new feat: commits in @qvac/cli currently at version 0.8.2. Following pre-major 0.x version policy, what exact version should be recommended?	pass→pass	6,770	3,945	-42%	1	1	0%	1,139	2,521	+121%	0	0	—
case-11 If @qvac/cli is updated to 0.9.0, will a dependent package specifier of ^0.8.0 match this new version automatically?	pass→pass	6,244	4,550	-27%	1	1	0%	1,111	2,527	+127%	0	0	—
case-12 When preparing a draft release PR for @qvac/cli version 0.9.0, what exact branch name must be created from main?	fail→pass	8,399	2,329	-72%	1	1	0%	1,430	2,159	+51%	0	0	—
case-22 Besides package.json, which plugin manifest file in plugins/openclaw requires a version update during release preparation?	fail→fail	6,188	1,944	-69%	1	1	0%	1,059	2,109	+99%	0	0	—
case-13 When executing release cascade preparation for @qvac/ai-sdk-provider, what secondary pull request must be created in the same session?	fail→fail	7,542	2,160	-71%	1	1	0%	1,194	2,182	+83%	0	0	—
case-14 Which git command flag must be included when cherry-picking release commits into the backmerge pull request?	pass→pass	6,003	2,559	-57%	1	1	0%	958	2,135	+123%	0	0	—
case-15 When updating files for a @qvac/cli release, what tool command generates the updated NOTICE file?	fail→fail	11,551	1,833	-84%	1	1	0%	1,890	2,085	+10%	0	0	—
case-16 When preparing a release for @qvac/ai-sdk-provider, which field in package.json must be updated to specify the CLI dependency version?	fail→fail	7,328	1,606	-78%	1	1	0%	1,178	2,051	+74%	0	0	—
case-17 What command installs @qvac/openclaw-plugin@latest without audit or funding notices in a temporary directory during promote verification?	pass→pass	6,104	2,023	-67%	1	1	0%	1,083	2,171	+100%	0	0	—
case-18 Which local script must be run inside packages/cli to verify publish readiness before promoting?	fail→pass	8,655	1,807	-79%	1	1	0%	1,367	2,112	+54%	0	0	—
case-19 In addition to @qvac/opencode-plugin, which other agent plugin package must be tracked in the agent stack sync cascade?	fail→pass	6,921	2,162	-69%	1	1	0%	1,201	2,159	+80%	0	0	—
case-20 When running the planner, npm shows version 0.8.3 published but no corresponding git tag exists in remote. What action must be taken?	fail→pass	8,401	2,712	-68%	1	1	0%	1,414	2,283	+61%	0	0	—
case-21 Before planning or preparing a cascade, what git command should be run to fetch tags and main from the org remote?	pass→pass	4,411	1,837	-58%	1	1	0%	672	2,119	+215%	0	0	—
case-23 What essential notification must be stated at the end of every agent stack sync report regarding package publishing?	pass→pass	5,834	1,964	-66%	1	1	0%	791	2,150	+172%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 16 counted toward the lift figure. The other 7 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +22 percentage points is the difference between those two pass rates over the 16 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 30.4%
Pass rate with the skill: 52.2%
Cases attempted: 23
Cases counted toward lift: 16
Cases that regressed: 3
Skill version measured: v2
Run completed: 8/22/2026, 17:20 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-08

Prompt

We have releases pending across the agent stack. What order must draft releases be prepped starting from the core SDK down to downstream agent plugins?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run52

Live evalNeeds ≥20 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+129% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 6 days ago

Other measured skills in the registry, with their headline benchmark lift.