Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add tmj-90/gaffer --skill api-versioninggit clone --depth 1 https://github.com/tmj-90/gafferWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tmj-90/gaffer/api-versioning)<a href="https://agentmods.dev/skills/tmj-90/gaffer/api-versioning"><img src="https://agentmods.dev/badge/skills/tmj-90/gaffer/api-versioning/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tmj-90/gaffer/api-versioning"><img src="https://agentmods.dev/badge/skills/tmj-90/gaffer/api-versioning.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00097 | $0.01225 |
| Opus 5.5 | $0.00039 | $0.00490 |
| Sonnet 5.5 | $0.00019 | $0.00245 |
| Haiku 4.5 | $0.00010 | $0.00122 |
Grade A, and why
api-versioning scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 16d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Version an API change
API clients are code you do not control, running versions you cannot see. Change compatibly by default: add, do not remove; widen what you accept, do not narrow what you return. When a break is unavoidable, ship a new version beside the old one, announce the deprecation in machine-readable headers, and remove the old version only on a published date.
Steps
- Classify the change against Google AIP-180 and the Microsoft REST guidelines.
Compatible: a new endpoint; a new optional request field whose absence keeps the old
behaviour; a new response field; accepting more input. Breaking:
- removing or renaming a field, endpoint, enum value or query parameter;
- changing a type, format, unit, or the serialisation of a value;
- making an optional request field required, or adding a required one;
- changing a default, a status code, an error code or the error body shape;
- no longer populating a field that used to be populated;
- same shape, different meaning (semantic break). A new value in a response enum breaks clients with exhaustive switches unless the contract already says the enum is extensible. When in doubt, it is breaking.
- Find the consumers. Call
search_lorefor the versioning scheme and declared consumers, andfind_dependentsfor the declared cross-repo consumers of the contract (an empty result is not proof of safety). Use the Grep tool for in-repo callers, generated clients and contract tests. Record who is affected. - Take the compatible path when one exists. Add the new field beside the old and populate both; accept old and new inputs; keep old defaults. Most "we need v2" tickets are an additive change plus a deprecation.
- If you must break, use the repo's one versioning mechanism (URL major version
/v2/, a header, or a datedapi-versionparameter); never add a second scheme. The versions share one domain implementation with a thin adapter per version at the edge, not a copied handler tree. An unsupported version gets a clear error listing the supported ones, in the repo's envelope (e.g.400for a badapi-versionvalue,404for an unknown/vN/prefix). - Pin every live version with contract tests (the
contract-testskill): the old version's response is asserted unchanged (field names, types, status codes, error body), and the new version's shape is asserted. Update the OpenAPI document for each version (theopenapi-contractskill). - Deprecate in machine-readable form. On every response from the deprecated
version or endpoint:
Deprecation: @<unix-seconds>(RFC 9745),Sunset: <HTTP-date>(RFC 8594), andLink: <doc-url>; rel="deprecation". Log each deprecated call with the caller identity so stragglers can be found; add a CHANGELOG entry. When external consumers exist, get the removal date throughrequest_decision; do not invent one. - Migrate the in-repo consumers you own in the same epic; a version no internal client uses is untested.
- Retire only on the date, via the
deprecate-and-removeskill: usage confirmed zero, the old version answers410 Gonewith a pointer to the replacement, then the code is deleted. - Evidence with the
record-evidenceskill: the classification and why, the consumer list, both contract tests passing, and a sample response showing the deprecation headers.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 16d ago Changed · +22 lines eea255c298bd
- 19d ago First seen · 60 lines · 97 tokens per session scan A b4b1a51c7a33
api-versioning is a skill published in the GitHub repository tmj-90/gaffer (2 stars, last pushed 17d ago), licensed Apache-2.0. It adds 97 tokens to every session and 1,225 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-28.
Other skills, from other repositories
bulwark-brainstorm
Role-based brainstorming with dual modes: --scoped (sequential Task tool, 5 roles) and --exploratory (Agent Teams peer debate, 4 roles). Use for feasibility assessment and idea validation.
plan-creation
Create structured implementation plans via a 4-role scrum team (Product Owner, Architect, Eng/Delivery Lead, QA/Critic) with optional Agent Teams peer debate mode.
anthropic-validator
Validates Claude Code assets (skills, hooks, agents, commands, MCP servers, plugins) against official Anthropic standards. Fetches latest docs dynamically and produces structured validation reports.
create-subagent
Generates single-purpose Claude Code sub-agents for use via the Task tool. Use when creating dedicated sub-agents, scaffolding agent definitions, or generating agents with diagnostics and permissions setup.
test-audit
Audit test suites for T1-T4 violations using AST analysis, mock detection, and multi-stage synthesis. Invoke when user asks to audit tests, check test quality, find mock violations, review test effectiveness, or inspect test suites for over-mocking. Triggers automatic rewrites when quality gates fail.
code-review
Comprehensive code review with distinct aspect based sections. Use when reviewing code, checking for security issues, finding type safety problems, auditing code quality, or when user asks to review code, PRs or changes. Three-phase workflow runs static tools, LLM judgment, and writes diagnostic log.