v0.4.7 — CITATION.cff and formal software citation metadata

What's Changed

• Root CITATION.cff and README Citing section for formal software citation (#269, #270)
• Maintainer release checklist for keeping citation version/date in sync

First Zenodo-archived snapshot after GitHub–Zenodo linking. Prefer the concept DOI once shown on Zenodo; record version 0.4.7 for reproducibility.

Full changelog: https://github.com/ARPAHLS/skillware/blob/main/CHANGELOG.md

v0.4.6 — card ui_schema CI, UK Companies House v2a, OpenAI-compatible hosts, wallet txlist pagination

Added

• Framework / CI — skillware/core/ui_schema.py and parametrized tests/test_card_ui_schema.py so every output-card ui_schema.fields[].key resolves against execute() fixtures (#199, #260).
• Documentation / Examples — OpenAI-compatible host guide (docs/usage/openai_compatible.md) plus runnable Groq example examples/openai_compatible_host.py — reuse to_openai_tool() with host base_url (Kimi, GLM, Groq, OpenRouter, …) (#261, #263).

Fixed

• compliance/mica_module, compliance/pii_masker — card.json output fields aligned with real execute() return shape (#199).

Changed

• finance/uk_companies_house_handler — Phase v2a: session context carry-forward, partial / pipeline prep, officer company_name fallback, interactive Gemini example (#220, #255).
• finance/wallet_screening — Paginate Etherscan txlist (up to 10k txs), surface metadata.warnings when truncated/unavailable; skill version 1.0.0 → 1.0.1 (#214, #215; parent #115).
• Documentation — CONTRIBUTING / TESTING / workflow: keep card.json and fixtures in sync when output shape changes (#199).

Upgrade

pip install -U skillware

Dev / multi-skill:

pip install -U "skillware[dev,all,agents]"

OpenAI-compatible host example (Groq):

pip install -U "skillware[compliance_tos_evaluator,openai]"
# set GROQ_API_KEY and GROQ_MODEL, then:
python examples/openai_compatible_host.py

See openai_compatible.md and CHANGELOG.md.

Contributors

• @rosspeili — card ui_schema validation guard and drift fixes (#199, #260); release cut
• @Areen-09 — UK Companies House handler v2a context carry-forward (#220, #255)
• @TheDarkniteFalls — OpenAI-compatible hosts guide and Groq example (#261, #263)
• @Hendobox — wallet_screening Etherscan txlist pagination and warnings (#214, #215)

Full changelog

• #260 — Audit registry card.json ui_schema keys against execute() output (#199)
• #255 — UK Companies House handler v2a (#220)
• #263 — Document OpenAI-compatible model hosts (#261)
• #215 — Paginate Etherscan txlist and surface warnings (#214)

Compare: v0.4.5...v0.4.6

v0.4.5 — creative/bg_remover, per-skill install extras, skillware paths

Added

• Skill — creative/bg_remover — offline background removal with rembg; base64 and file I/O, transparent PNG output, mocked offline tests (#196, #244). First skill in the new creative category.
• Packaging — Category, per-skill, and [all] optional extras generated from skill manifests via scripts/sync_extras.py; hand-maintained dev, SDK extras, and meta [agents] (#236).
• Documentation — Install extras — canonical guide for pip install skillware[...]; every skill catalog page recommends its per-skill extra (#236).
• Framework — skillware/core/extras.py and richer ImportError hints from SkillLoader when manifest requirements are missing (#236).
• Framework — BaseSkill.validate_params() — optional JSON Schema validation helper for tool arguments (#125).
• Framework — Registry manifests standardize on outputs: (legacy singular output: removed from finance/wallet_screening) (#125).
• CLI — skillware paths — skill root resolution order (external → project → bundled), tier labels, shadowing summary, operator tips; interactive menu option 4 (#81).
• Framework — skillware/core/discovery.py — shared skill root discovery for SkillLoader and the CLI (#81).
• CI — PyPI wheel packaging smoke test — builds a wheel, installs in a fresh venv (base deps only), verifies every bundled registry skill is loadable via scripts/wheel_smoke_test.py (#182).

Changed

• Packaging (breaking) — Removed legacy extras cli and embeddings. [all] is skill runtime deps only (not SDK packages). Use [agents] for Gemini + Claude + OpenAI SDK deps (#236).
• CLI — skillware list prints a pointer to the install extras guide (#236).
• Framework — SkillLoader delegates root resolution to discovery.get_skill_roots() so load, list, test, and paths share the same order (#81).
• Framework — Map manifest requirement pillow to import name PIL in dependency checks (#196).
• Documentation — README, CONTRIBUTING, TESTING, skill catalog, examples index updated for the new extra taxonomy (#236).
• Documentation — Skill trust model operator expectations and instruction-only content notes (#243).

Upgrade

pip install -U skillware

New skill (background removal):

pip install -U "skillware[creative_bg_remover]"

Full local dev / multi-skill installs:

pip install -U "skillware[dev,all,agents]"

Per-skill extras (recommended in docs even when empty today):

pip install "skillware[finance_wallet_screening]"
pip install "skillware[office_pdf_form_filler]"

Inspect skill roots after upgrade:

skillware paths
```bash

See [install_extras.md](https://github.com/ARPAHLS/skillware/blob/main/docs/usage/install_extras.md) and [CHANGELOG.md](https://github.com/ARPAHLS/skillware/blob/main/CHANGELOG.md#045---2026-07-16).

## Contributors

- **@AyushSrivastava1818** — `creative/bg_remover` skill bundle, tests, and catalog docs (#196, #244)
- **@rosspeili** — Per-skill install extras and manifest sync (#236); wheel packaging smoke test (#182); `validate_params()` and `outputs:` standardization (#125); `skillware paths` and discovery module (#81); maintainer finish on #244; release cut
- **@mrmasa88** — Skill trust model operator expectations and instruction-only content notes (#243, #250)

## Full changelog

- #244 — `creative/bg_remover` offline background removal skill (#196)
- #256 — Category and per-skill pip extras with manifest sync (#236)
- #254 — PyPI wheel packaging smoke test for bundled registry skills (#182)
- #253 — Manifest param validation and `outputs:` standardization (#125)
- #250 — Skill trust model operator expectations (#243)
- #249 — `skillware paths` and shared discovery module (#81)

**Compare:** https://github.com/ARPAHLS/skillware/compare/v0.4.3...v0.4.5

v0.4.2 — CLI splash polish, architecture diagrams, docs sweep

Changed

• CLI — Interactive splash: corrected SKILLWARE ASCII logo, Rich 3-stop gradient on logo and tagline, shared _package_version_str() for splash and skillware --version, tagline Skillware v{version} — Skill Management Framework (#222, #224).
• Documentation — Compact architecture Mermaid diagrams: shorter README labels, model-agnostic adapter nodes in introduction, horizontal agent-loop layout with role table (#210, #217, #221).
• Documentation — Docs sweep: README Registry Mermaid alignment, vision.md pinned to v0.4.x, skill catalog and usage guides prefer bundle["class"]() with explicit bundle["module"].ClassName() noted as still supported (#225).

Upgrade

pip install -U skillware

Interactive splash and CLI unchanged for programmatic use (SkillLoader, agent loops). After upgrade:

skillware

Dev checkout:

pip install -e ".[dev,all]"

See cli.md and CHANGELOG.md.

Contributors

• @rosspeili — CLI splash gradient and version alignment (#222, #224); compact diagram polish (#221); docs sweep — vision, bundle class usage, README Mermaid (#225)
• @meghana922007 — Architecture flow diagrams (#217, #210)

Full changelog

• #225 — Docs sweep: vision v0.4.x, bundle["class"]() in catalog and usage guides, README Registry alignment
• #224 — CLI interactive splash: SKILLWARE logo, Rich gradient, shared version string (#222)
• #221 — Compact architecture Mermaid diagrams (#210 follow-up)

Compare: v0.4.1...v0.4.2

v0.4.1 — uk_companies_house_handler, loader improvements, trust model, architecture diagrams

Added

• finance/uk_companies_house_handler — Deterministic UK Companies House REST integration: company search, profile, officers, PSC, filing history, intent-to-operation mapping, UK terminology translation, and status envelope (ready/needs_input/error) with disambiguation (#172, #218).
• Documentation — Skill trust model (docs/security/skill-trust-model.md): execution model, resolution order, Bundled / Project / External provenance tiers, operator guidance; wired from SECURITY, usage, CONTRIBUTING, and CODE_OF_CONDUCT (#109).
• Documentation — Mermaid architecture diagrams in README, introduction, and agent loops; cross-links, Step 1 mini-pipeline, adapter fan-out, and direct-path footnotes (#210, #217).

Changed

• Loader — Auto-discover BaseSkill subclass as bundle["class"]; add get_skill_class(). bundle["module"] unchanged (#89).
• Loader — Warn when registry manifest.yaml name diverges from path; expose optional registry_id (#200).
• Version policy — Security support floor >= 0.3.5; legacy 0.3.0–0.3.4 silent; unsupported advisory below 0.3.0 (#192).
• Documentation — Clarify skill ID vs manifest name vs provider tool names (#201).

Fixed

• office/pdf_form_filler and defi/evm_tx_handler — Align manifest name with registry paths; update examples and docs for manifest-derived tool dispatch (#201).

Upgrade

pip install -U skillware

finance/uk_companies_house_handler needs COMPANIES_HOUSE_API_KEY (see .env.example and api_keys.md).

from skillware.core.loader import SkillLoader

SkillLoader.load_skill("finance/uk_companies_house_handler")

See uk_companies_house_handler.md and examples/README.md.

Contributors

• @Areen-09 — finance/uk_companies_house_handler skill, bundle, examples, and catalog docs (#218)
• @mrmasa88 — Skill trust model documentation (#212, #109)
• @meghana922007 — Architecture flow diagrams (#217, #210)
• @rosspeili — Loader class discovery and identity validation (#89, #200); version policy (#192); manifest ID alignment (#201)

Full changelog

• #218 — UK Companies House handler (#172)
• #212 — Skill trust model (#109)
• #217 — Architecture diagrams (#210)
• #216 — Loader bundle["class"] discovery (#89)
• #213 — Manifest name / registry ID validation (#200)
• #211 — Version policy thresholds (#192)
• #209 — Manifest name alignment and docs (#201)

Compare: v0.4.0...v0.4.1

v0.4.0 — monitoring/token_limiter, doc-drift CI guards, README stats

Added

• monitoring/token_limiter — Deterministic token budget gate for autonomous agent loops. Host passes cumulative token usage; skill returns CONTINUE, WARN, or FORCE_TERMINATE. Bundled indicative model pricing, ROI scaffold fields for v2 (not enforced), catalog docs, and local + Gemini/Claude loop examples (#23, #207).
• New monitoring skill category in CONTRIBUTING, ai_native_workflow.md, and Skill Library.
• Tests — tests/test_registry_docs.py CI doc-drift guards: skill catalog, examples/README.md, and agent_loops.md stay in sync with the registry (#183, #189).
• Documentation — Cross-linked doc-drift guards in contributor guides; PR checklist reminder in CONTRIBUTING (#193, #197).

Changed

• Documentation — token_limiter catalog page: budget disclaimer, limitations, enterprise disclaimer (#23).
• Documentation — README Stats section and live PyPI download badges (#198).
• Documentation — CLI and examples docs aligned with skillware test / skillware examples behavior (#191, #194).

Upgrade

pip install -U skillware

monitoring/token_limiter has no extra runtime dependencies. Load with:

from skillware.core.loader import SkillLoader

SkillLoader.load_skill("monitoring/token_limiter")

See token_limiter.md and examples/README.md (token_limiter_loop.py, gemini_token_limiter.py, claude_token_limiter.py).

Contributors

• @rosspeili — monitoring/token_limiter skill, bundle, examples, and catalog docs (#23, #207); README Stats badges (#198)
• @Areen-09 — tests/test_registry_docs.py doc-drift guards (#189); contributor guide cross-links (#197)
• @uttam12331 — CLI and examples documentation alignment (#194)

Full changelog

• #207 — monitoring/token_limiter skill (#23)
• #189 — Documentation consistency tests (#183)
• #194 — CLI/examples docs alignment (#191)
• #197 — Doc-drift guard cross-links (#193)
• #198 — README Stats and PyPI download badges

Compare: v0.3.9...v0.4.0

v0.3.9 — examples CLI loads index from GitHub for pip installs

Fixed

• CLI — skillware examples and skillware list --examples fetch examples/README.md from GitHub when no local copy exists (typical after pip install). Local checkouts still use the repo file first.

Upgrade

pip install -U skillware

The CLI (skillware list, skillware test, skillware examples, interactive menu) is included in the core install. skillware examples needs network access when run outside a Skillware checkout.

Contributors

• @rosspeili — GitHub fallback for examples index; remove unused wheel packaging of examples/README.md

Full changelog

• #190 follow-up — Fetch examples/README.md from GitHub for pip-only installs (0.3.8 regression fix)

Compare: v0.3.8...v0.3.9

v0.3.8 — skillware test & examples CLI, bundle test enforcement, testing docs

Added

• CLI — skillware test runs bundle tests via pytest (all roots, by skill ID, or --category); supports -v and --no-header. Documented in cli.md, TESTING.md, and contributor guides (#83, #188).
• CLI — skillware list --examples shows per-skill example script counts; skillware examples [skill_id] lists indexed runnable scripts from examples/README.md with GitHub source links; interactive menu option examples. Ships examples/README.md in the wheel so pip installs can resolve the index (#126, #190).

Changed

• Tests — tests/test_skill_issuer.py now requires test_skill.py for every registry skill under skills/ (#160, #185).
• Documentation — Clarified that bundle tests must mock network calls and model downloads in CI (#170, #184).
• Documentation — Added a Status section to TESTING.md summarizing the current testing model (#179, #186).
• Documentation — Post-release alignment — category tables, Python 3.10+ badge, dev install ([dev,all] vs [dev]), README configuration via .env.example, DeFi env vars in .env.example, framework env vars in api_keys.md (#154, #187).

Upgrade

pip install -U skillware

The CLI (skillware list, skillware test, skillware examples, interactive menu) is included in the core install.

Contributors

• @rosspeili — skillware test CLI (#83, #188); skillware examples and list --examples (#126, #190); bundle test enforcement (#160, #185); TESTING.md status section (#179, #186); post-release docs alignment (#154, #187)
• @katebabakova444 — Bundle test mocking guidance in CONTRIBUTING and TESTING.md (#170, #184)

Full changelog

• #184 — Clarify bundle tests mock network and model downloads (#170)
• #185 — Enforce test_skill.py for every registry skill (#160)
• #186 — TESTING.md Status section (#179)
• #187 — Post-release documentation alignment (#154)
• #188 — skillware test CLI for bundle tests (#83)
• #190 — skillware examples CLI and list --examples (#126)

Compare: v0.3.7...v0.3.8

v0.3.7 — wellness/mental_coach, maintainer test layout, wallet_screening doc

Added

• wellness/mental_coach — Deterministic wellness coaching firewall with crisis triage, hard scope limits, embedded public KB, optional Gemini scope evaluator, catalog documentation, and mental_coach_demo.py example (#148, #174).
• New wellness skill category in CONTRIBUTING and docs/usage/agent_loops.md.

Changed

• Tests — Moved tests/test_mica_module.py to tests/skills/compliance/test_mica_module.py so maintainer skill tests follow the tests/skills/<category>/ layout; tests/ root is framework-only (#86, #180).

Fixed

• finance/wallet_screening — Align examples and docs with finance/wallet_screening manifest tool name; fix gemini_wallet_check.py and claude_wallet_check.py to match tool name dynamically from manifest; correct card.json UI fields to match actual skill output schema; update instructions.md, provider snippets, Data Schema, and usage docs (#173, #175).

Upgrade

pip install -U skillware

Optional skill runtime deps (including wellness/mental_coach evaluator): pip install "skillware[all]".

Contributors

• @mrmasa88 — wellness/mental_coach skill: crisis gate, grounded KB, bundle tests, catalog page, and demo (#148, #174)
• @rosspeili — Maintainer test layout: move mica_module test under tests/skills/compliance/ (#86, #180)
• @Akanksha-H — finance/wallet_screening examples, docs, and card.json alignment with manifest tool name (#173, #175)

Full changelog

• #174 — Add wellness/mental_coach skill resolving #148
• #175 — wallet_screening examples and docs aligned with manifest tool name (#173)
• #180 — Move mica_module maintainer test under tests/skills/compliance/ (#86)

Compare: v0.3.6...v0.3.7

v0.3.6 — CLI discoverability, bundle tests in CI, wallet_screening manifest fix

Added

• Tests — Backfilled test_skill.py for six registry skills (mica_module, pii_masker, synthetic_generator, wallet_screening, pdf_form_filler, prompt_rewriter); all registry skills now ship co-located bundle tests. Fixed prompt_rewriter package export so pytest can collect the bundle (#158).
• CLI — skillware/__main__.py enables python -m skillware as a fallback when the skillware command is not on PATH (#135). Added cmd_help() for rich-formatted help, wired to skillware --help/-h and interactive menu option 4. Added --version/-V flag.

Fixed

• novelty_extractor — Bundle tests mock fastembed embeddings so CI avoids HuggingFace downloads and rate limits (#159 follow-up).
• finance/wallet_screening — WalletScreeningSkill.manifest loads manifest.yaml from the bundle directory (#165).

Changed

• CI — GitHub Actions runs pytest skills/ then pytest tests/ after lint (bundle + framework/maintainer tests; closes #90) (#159).
• CI — CodeQL GitHub Action upgraded from v3 to v4.
• Dependencies — Extended [all] with registry skill runtime deps (web3, fastembed, numpy); added [defi] and [embeddings] optional extras. Documented manifest ↔ pyproject.toml convention in CONTRIBUTING and TESTING.md.
• Documentation — TESTING.md, CONTRIBUTING.md, ai_native_workflow.md, and README architecture tree document the bundle / framework / maintainer / example testing model. Pytest collects tests/ and skills/ only (examples/ ignored).
• Dependencies — Moved rich>=13.0 from [cli] extra to core dependencies; CLI is now available immediately after pip install skillware. The [cli] extra is kept as an empty deprecated alias for backward compatibility (#135).

Upgrade

pip install -U skillware

The CLI (skillware list, interactive menu, python -m skillware) is included in the core install — no [cli] extra required.

Contributors

• @rizzoMartin — CLI discoverability: python -m skillware, unified rich help, --version, rich in core (#135)
• @Hendobox — finance/wallet_screening manifest loading fix (#165)
• @rosspeili — Bundle test backfill (#158), CI pytest skills/ (#159), novelty_extractor CI mock follow-up

Full changelog

• #135 — CLI discoverability, PATH docs, python -m skillware, unified help (#147)
• #158 — Backfill bundle tests for six registry skills (#164)
• #159 — CI runs pytest skills/ alongside pytest tests/ (#168)
• #169 — novelty_extractor bundle tests mock fastembed for offline CI (#159 follow-up)
• #163 — CodeQL GitHub Action upgraded v3 → v4
• #165 — wallet_screening skill.manifest loads manifest.yaml (#171)

Compare: v0.3.5...v0.3.6