agentleFS
nesnilnehcnesnilnehc/ai-cortex/skills/bump-version/SKILL.md

bump-version

Skill8 starsChanged 2 months agoMITView on GitHub

Update a project's version references to an explicit target version or calculate the next version from a supplied increment, then verify consistency. Use for mechanical version changes; compatibility analysis, release scope, changelog authoring, and delivery are separate tasks.

42 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use oss-pr, pr-automation and 7 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • CES-LtdSkill

    oss-pr

    Use when creating a pull request, after committing changes, or when user invokes /oss-pr. Covers branch management, quality checks, commit, push, and PR creation.

    Also in 3 projects that use this

  • CES-LtdSkill

    pr-automation

    PR Automation Orchestrator: poll open PRs, check CI, run review, fix, and merge eligible PRs. Use when: (1) Invoked by daemon via scripts/pr-automation.sh, (2) User says "/pr-automation".

    Also in 3 projects that use this

  • volovelySkill

    pr-fix

    This skill should be used when the user asks to "fix PR comments", "address PR feedback", "resolve PR comments", provides a GitHub PR URL with intent to fix, or says "pr-fix". It reads all review comments on a PR, decides which need code fixes vs. replies, makes the fixes, and posts replies — all with explicit user approval before any GitHub action.

    Also in 3 projects that use this

  • lbednerSkill

    i18n

    Use when a code or template change introduces a new translation message key, in either the framework CLI (aegis/i18n/locales/) or a generated project's CLI (the app/i18n/locales/ template tree). Covers adding the real English string, stubbing the same key into every other locale for parity, and the separate translation PR that later replaces placeholders with reviewed native text.

    Also in 4 projects that use this

  • AnthropicSkill

    architecture

    Create or evaluate an architecture decision record (ADR). Use when choosing between technologies (e.g., Kafka vs SQS), documenting a design decision with trade-offs and consequences, reviewing a system design proposal, or designing a new component from requirements and constraints.

    Also in 4 projects that use this

  • NVIDIASkill

    testing

    Run Python and Rust tests for physicsnemo-curator using uv + pytest with coverage reporting via pytest-cov, and cargo-nextest for Rust tests. Includes benchmark workflows with pytest-benchmark and criterion.

    Also in 5 projects that use this

  • +2 more from CES-Ltd

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 39 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read project instructions.

    line 22
  2. Distinctive:

    If the target already matches all required references, report no change.

    line 36

“Common” is measured: kept by at least 6 in 10 of the 39 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Bump Version

Apply an already chosen version or increment through the project's version mechanism and verify the resulting references. Support different projects without hardcoding package managers, file names, registries, or Git conventions. Calculating a version from a supplied increment is arithmetic under the project's rules, not a recommendation about the impact of code changes.

Inputs

Require a project/component and either an exact target version or an explicit increment supported by its version policy. The current checkout can supply project context. Discover current version and policy from authoritative project configuration and instructions.

The version decision may come from the user, an offline decision or an existing project record; no version-planning skill is required. Version edits and separately prepared change descriptions may be revised in either order without forcing one capability to invoke the other. This skill verifies version references; description facts and scope, code/package verification, and reviewed commits belong to their respective project mechanisms. A version edit does not require an automatic commit.

If neither target nor increment is available, ask for that specific input before editing. Do not inspect requirements, commit history, a release diff, or a registry's latest version to choose major/minor/patch. If the user asks which increment is warranted, explain the missing impact decision and leave that analysis to a separate task. An exact target fixes the intended value; if both target and increment are supplied and disagree, resolve the conflict rather than choosing one silently.

Discover the version mechanism

  1. Read project instructions. Inspect repository root, branch, HEAD, staged/unstaged changes, and existing version-update scripts/configuration. Preserve unrelated work and its staging.
  2. Identify the selected component and version domain, authoritative current version, version scheme, synchronized references, and project-owned update entry point. Product, API, protocol and data-format versions can be independent, including within a single component. In a monorepo, establish whether versions are independent or coordinated. Update only the selected domain and components plus those the project's rules require to move with them. If the selection or required coordination is ambiguous, ask before editing.
  3. Before executing any project command, including consistency checks and other verification commands, inspect its invoked commands and lifecycle hooks for changelog generation, builds, Git commits/tags/pushes, publication, or other effects outside this task. Disable these effects only through a documented scoped mode or use a narrow supported file edit. For verification, use direct comparison of required references when the command cannot safely run. If safety cannot be established or no safe method is available, report the gap rather than running the broader workflow.

Prefer existing project tools and available local dependencies. Do not add a version-management dependency or global installation merely to increment a version. Existing lockfiles should be updated by their package manager where supported, not by broad text replacement. If the necessary tool is unavailable, report what is missing and preserve the current state.

If a tag or build-time generator is the sole version authority, report how this target would be supplied through that mechanism. Do not invent a persistent version file or create a tag. If the project has no established version mechanism, identify the missing project decision rather than designing one inside this operation.

Resolve the exact target

  • For an explicit target, validate its syntax and applicable project constraints. Do not reclassify its compatibility impact.
  • For an explicit increment, calculate the target from the authoritative current version using the project's scheme and supported tool. SemVer projects follow their SemVer arithmetic; calendar or custom versions follow their own documented rules. Do not impose SemVer or infer a new numbering policy.
  • A prerelease transition needs the project's existing rule and enough input to resolve the identifier/counter; do not guess the channel or reset counters silently.
  • If sources disagree or an unfinished version edit exists, inspect ownership and the intended target. Resume a clearly identified change at the same target rather than incrementing it again. If an increment is repeated after a completed update, do not assume it is a retry or a new request: resolve the intended target from the request and available evidence, and ask if ambiguous. If unresolved, report the specific conflict before modifying files.
  • If the target already matches all required references, report no change. Reject an unsupported decrement or reuse under the project's rules; an explicit correction permitted by those rules is not automatically a release.

An availability check in a registry, published-artifact baseline, or last version-bump commit is not required for this local edit. Publication/version-allocation checks belong to the relevant release or delivery workflow. Do not claim a locally valid version is available to publish.

Before editing, fix the baseline, exact target and required synchronization mapping for each selected component/version domain in the working context. On interruption or partial failure, retain those targets and complete only the outstanding references; do not calculate another increment from an already updated value. For example, if a 1.2.3 → 1.3.0 change updates one of two required references before interruption, resume by bringing the remaining reference to 1.3.0, not by advancing to 1.4.0. Recover the intended targets from available evidence when needed; if they remain ambiguous, ask before editing. This does not require a new persistent state file.

Modify and verify

Apply the exact target only to the selected version domain and required synchronized references and generated metadata. An explicitly selected API, protocol or data-format version is a valid target. Change dependency constraints or other version domains only when the project's rules require them to synchronize with that target; do not change unrelated versions, application behavior, requirements, or historical release entries merely because they contain similar numbers.

Run the project-supported consistency check, or explicitly compare the authoritative value and every reference that must synchronize according to the project's mapping; coordinated domains need not have identical version numbers. Missing or ambiguous references are failures, not default values. Inspect the diff for unintended edits and check whitespace. Run additional targeted checks only when the project requires them or the version change affects generated code or behavior; a full test/build/package run is not an automatic requirement for a metadata-only change.

On failure, keep the edits inspectable, report which values/files differ, and fix only within the authorized scope. Do not reset or stash unrelated work, weaken checks, or calculate another target to conceal a failed attempt. Reuse prior passing evidence only while its relevant inputs still match.

Output and boundary

Report the previous and exact target versions, whether the target was supplied or calculated from a supplied increment, the project mechanism, changed file links, consistency-check results, and remaining blockers. If the version is derived solely from tags or build parameters, explain that no persistent file change was made and give the established handoff. Keep the response short and distinguish verified equality from merely displaying values.

This skill ends with verified local version edits (or an explained no-op/blocker). Do not select release scope, analyze compatibility, write changelogs/release notes, commit, create/move tags, push, build deliverables, or publish. Those are separate operations; a request to bump a version alone does not authorize them. No companion skill or specific Agent API is required to perform this capability.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.