chore(seo): retarget PyPI metadata for category-search discovery (#56)

* chore(seo): retarget PyPI metadata for category-search discovery

The Python adoption side has been lagging TypeScript materially (PyPI
flat for ~6 weeks pre-source-change; cycles-client-python at 157 unique
GitHub cloners vs cycles-client-typescript at 288). Diagnostic showed
the biggest sub-gap is discovery: Cycles doesn't rank for natural
Python search queries while named competitors (AgentBudget, AgentGuard,
agent-cost-guardrails) do. The brand-led PyPI description was a
contributor.

Three changes:

1. Description rewritten to lead with the literal category-search
   phrase "Python AI agent budget control" instead of the brand-led
   "Python client for the Cycles budget-management protocol."

2. Keywords expanded from 12 to 21, organized into three groups:
   category-search keywords (ai-agent, agent-budget, budget-control,
   cost-control, cost-enforcement, spending-limit, llm-cost,
   runtime-authority, action-control, multi-tenant), framework
   targeting (langchain, langgraph, crewai, autogen, openai-agents,
   mcp, openai, anthropic), and brand (cycles, runcycles). Drops
   `metering` (low search volume) and adds the missing framework tags.

3. Classifier added: Topic :: Scientific/Engineering ::
   Artificial Intelligence — standard PyPI classifier for AI/ML
   packages, materially improves browse-by-category discovery.

Takes effect on next PyPI publish (version bump required). Suggest
0.4.0 → 0.4.1 since this is metadata-only.

Companion change: GitHub topics already updated separately
(governance dropped, mcp added) to mirror cycles-client-typescript's
topic set.

* chore: bump version to 0.4.1 + CHANGELOG / AUDIT entries

Metadata-only release. Bumps pyproject.toml version 0.4.0 → 0.4.1
and adds the corresponding CHANGELOG and AUDIT.md entries for the
PyPI discovery-metadata refresh shipped earlier in this branch.

No code, no test, no protocol changes — wire format and public API
are identical to 0.4.0. Ready for `python -m build` + `twine upload`
on merge.

Author: amavashev

AUDIT.md

@@ -224,3 +224,18 @@ Widened the `@cycles` decorator to accept callables — in addition to constants
 - **Signature change:** `_build_reservation_body` now takes `args` and `kwargs` parameters; both `CyclesLifecycle.execute` and `AsyncCyclesLifecycle.execute` thread them through.
 
 Protocol conformance: No protocol or wire-format changes. The reservation request body shape is unchanged — only the source of each field's value is widened. Verified by new unit tests in `TestCallableSubjectFields`, `TestCallableActionFields`, `TestCallableDimensions` plus an end-to-end decorator test asserting the captured request body.
+
+## PyPI Metadata Refresh (added 2026-05-08)
+
+**Files:** `pyproject.toml`
+**Version:** 0.4.1
+
+Metadata-only release retargeting the package for category-search discovery on PyPI. No code, no test, no protocol changes — wire format and API are identical to 0.4.0.
+
+- **Description rewritten** to lead with the literal category-search phrase: `"Python AI agent budget control — enforce LLM cost limits, tool permissions, and multi-tenant policies before agent actions execute."`
+- **Keywords expanded 12 → 21**, organized into category-search terms (`ai-agent`, `agent-budget`, `budget-control`, `cost-control`, `cost-enforcement`, `spending-limit`, `llm-cost`, `runtime-authority`, `action-control`, `multi-tenant`), framework targeting (`langchain`, `langgraph`, `crewai`, `autogen`, `openai-agents`, `mcp`, `openai`, `anthropic`), and brand (`cycles`, `runcycles`).
+- **Classifier added:** `Topic :: Scientific/Engineering :: Artificial Intelligence`.
+
+Driven by Python-side adoption diagnostic finding the biggest sub-gap was discovery, not SDK feature parity. Companion changes: GitHub topics on this repo (`governance` dropped, `mcp` added) and Python framework integration guide retitling on `runcycles/docs` (PR #568).
+
+Protocol conformance: No protocol or wire-format changes. Existing test suite at 100% coverage; no test additions.

CHANGELOG.md

@@ -5,6 +5,14 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog 1.1.0](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.4.1] - 2026-05-08
+
+PyPI metadata refresh for category-search discovery. No code changes; package wire format and API are identical to 0.4.0.
+
+### Changed
+
+- `pyproject.toml`: rewrote `description` to lead with the literal category-search phrase ("Python AI agent budget control — enforce LLM cost limits, tool permissions, and multi-tenant policies before agent actions execute"). Expanded `keywords` from 12 to 21, organized into category-search terms, framework targeting (`langgraph`, `crewai`, `autogen`, `openai-agents`, `mcp`, etc.), and brand. Added `Topic :: Scientific/Engineering :: Artificial Intelligence` classifier for PyPI browse-by-category surfacing.
+
 ## [0.4.0] - 2026-04-27
 
 Dynamic subject and action fields on the `@cycles` decorator.
@@ -120,6 +128,7 @@ Initial public release.
 
 - Comprehensive error handling and improved API model validation (#1)
 
+[0.4.1]: https://github.com/runcycles/cycles-client-python/compare/v0.4.0...v0.4.1
 [0.4.0]: https://github.com/runcycles/cycles-client-python/compare/v0.3.0...v0.4.0
 [0.3.0]: https://github.com/runcycles/cycles-client-python/compare/v0.2.0...v0.3.0
 [0.2.0]: https://github.com/runcycles/cycles-client-python/compare/v0.1.3...v0.2.0

pyproject.toml

@@ -4,23 +4,35 @@ build-backend = "hatchling.build"
 
 [project]
 name = "runcycles"
-version = "0.4.0"
-description = "Python client for the Cycles budget-management protocol"
+version = "0.4.1"
+description = "Python AI agent budget control — enforce LLM cost limits, tool permissions, and multi-tenant policies before agent actions execute."
 readme = "README.md"
 license = "Apache-2.0"
 requires-python = ">=3.10"
 keywords = [
-    "cycles",
-    "budget",
+    # Category-search keywords (lead with these — what people actually type)
+    "ai-agent",
+    "agent-budget",
+    "agent-governance",
+    "budget-control",
     "cost-control",
-    "metering",
-    "ai",
-    "llm",
-    "agents",
-    "governance",
+    "cost-enforcement",
+    "spending-limit",
+    "llm-cost",
+    "runtime-authority",
+    "action-control",
+    "multi-tenant",
+    # Framework-specific (the integrations we want to surface for)
+    "langchain",
+    "langgraph",
+    "crewai",
+    "autogen",
+    "openai-agents",
+    "mcp",
     "openai",
     "anthropic",
-    "langchain",
+    # Brand
+    "cycles",
     "runcycles",
 ]
 classifiers = [
@@ -32,6 +44,7 @@ classifiers = [
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
     "Topic :: Software Development :: Libraries",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
     "Framework :: AsyncIO",
     "Typing :: Typed",
 ]