@djordje-stojanovic/sigmaskills 0.16.7 → 0.16.9

package/CHANGELOG.md

@@ -2,8 +2,62 @@
 
 All notable changes to [Sigmaskills](https://github.com/Djordje-Stojanovic/Sigmaskills) are documented here.
 
+## Ticket 83 (2026-10-19 to 2026-10-23)
+
+66 merged pull requests and 68 commits on `main`, shipped as releases 0.10.0 to 0.16.8 (each has its own section in this file). The picture across all four repos: the [ticket 83 changelog page](https://claude.ai/artifact/U1Z7U3z1BMFVu3rUPXAdMJ).
+
+| Release PR | Merged (Vienna) | Title |
+|---|---|---|
+| [#119](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/119) | 10-19 03:53 | release: 0.10.0 - sigmaloop, sigmataste, sigmaship upgrade |
+| [#131](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/131) | 10-20 03:57 | release: 0.11.0 - sigmahuman, sigmaresearch persistence ladder, sigmaloop host files |
+| [#136](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/136) | 10-20 15:03 | Release identities for 0.12.0 |
+| [#139](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/139) | 10-20 17:36 | release: identities for 0.13.0 (sigmafork course v2) |
+| [#142](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/142) | 10-21 01:31 | release: identities for 0.14.0 (sigmatour, sigmaresearch Reddit/X routes) |
+| [#147](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/147) | 10-21 09:17 | release: identities for 0.15.0 (sigmafinanceman) |
+| [#150](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/150) | 10-21 16:50 | release: identities for 0.15.1 (sigmadeeplook --ask fix) |
+| [#152](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/152) | 10-21 18:09 | release: identities for 0.15.2 (Claude Code run mode) |
+| [#159](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/159) | 10-22 23:18 | release: identities for 0.16.0 (sigmadeeplook round 8 and cleanup) |
+| [#162](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/162) | 10-23 00:48 | release: identities for 0.16.1 (sigmadeeplook fix round 9) |
+| [#165](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/165) | 10-23 01:47 | release: identities for 0.16.2 (skill descriptions at 200 characters) |
+| [#167](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/167) | 10-23 11:52 | release: identities for 0.16.3 (sigmafinanceman masters and method) |
+| [#172](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/172) | 10-23 13:06 | release: identities for 0.16.4 (sigmadeeplook S13 round 2, sigmafinanceman FRED fallback) |
+| [#175](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/175) | 10-23 14:39 | release: identities for 0.16.5 (sigmadeeplook fix round 10) |
+| [#179](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/179) | 10-23 17:32 | release: identities for 0.16.6 (sigmadeeplook fix round 11) |
+| [#182](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/182) | 10-23 19:54 | release: identities for 0.16.7 (installer pack fix) |
+| [#185](https://github.com/Djordje-Stojanovic/Sigmaskills/pull/185) | 10-23 21:07 | release: identities for 0.16.8 (four new packs) |
+
 ## [Unreleased]
 
+## [0.16.9] — 2026-10-24
+
+### Fixed
+
+- `sigmadeeplook --ask` (fix round 12, lane F1, from held-out set 11):
+  - A big `.sav` is searched to the last case. A rounding number ("to 1 decimal", "top 5") no longer counts as a key, and a search that fills its keep list says PARTIAL instead of "0 matching".
+  - A number in quotes (`time index "3287"`, `pixel "336,519"`, curly quotes, backticks) reads as the plain number in netCDF and picture questions. `(x,y)` alone names a pixel on a picture.
+  - Step 1 of every table-like file has an `Ask a count:` line: the count forms `--ask` answers, with the file's own column names.
+  - `contain the word 'X'` is a real filter over cue text and text columns: whole word, case as written, `ignoring case` to ignore it.
+  - A count whose condition matches no filter now says `no filter matched` and lists what it can do. It no longer answers with the row total (the sbv "1,240 cues" answer). An id or code column is never summed for "how many".
+  - A question that writes a table's own name (`gpkg_contents`) is answered from that table, not from a look-alike (`gpkg_ogr_contents`).
+- `sigmadeeplook` fix round 12, lane F2 (set 11 wrong answers and print gaps):
+  - A frequent-word count now counts what `--ask` counts. The capitalised-word list counted a word joined to a digit or an underscore (`Table11`, `AB1234Capsules`) as the word, so it said 20 where the whole-word count is 19. The line now says it is a case-sensitive count of whole words.
+  - `--ask` no longer reads a quoted word as a table question just because the word is `Table`. A question on a paragraph count gives two figures when the letters also sit inside a longer word: `21 paragraphs contain 'Capsules'; 22 paragraphs hold the letters 'Capsules' (1 only inside a longer word, e.g. 'AB1234Capsules')`. An occurrence count adds the same clause. Plain `.txt`, `.md` and `.rst` files are counted too, with their own line numbers (a paragraph is a block between blank lines).
+  - Video: a new `Clock:` line gives the container start, every stream start and the first video frame. Black and frozen spans are told from the first video frame, with the container time beside them (`black from 0.0 s to 4.8 s after the first video frame (container time 1.44 s to 6.24 s)`), and the first non-black frame is named. Before, a span "at 0:00" counted from the container start, which is 1.4 s in front of the first frame of a transport stream.
+  - New `Heading styles` line for reST and Markdown: headings per underline character (`'-' underline x17`), over-and-under titles, and `#` marks. `--ask "how many headings are underlined with hyphens"` and `"... start with '##'"` answer it.
+  - A value that names the file is no longer cut: a path with one value in an XML file (the head title of an OPML file), the last record's fields, document properties, container tags and the first and last row of a table print whole up to 200 characters. A longer value says `(cut, N chars)` and the print says how to get the rest.
+  - BMP, PCX and TGA headers state the stored bytes per row with padding, and `--ask "how many bytes does one stored pixel row occupy"` answers it (BMP rows are padded to a multiple of 4 bytes). TGA now has a header line.
+
+## [0.16.8] — 2026-10-23
+
+### Changed
+
+- Four new packs, each pinned to one commit or tag: `anthropic` (19 skills from anthropics/skills), `design-md` (84 brand looks from VoltAgent/awesome-design-md, one skill per brand), `last30days` (mvanhorn/last30days-skill v3.30.0) and `financial-services` (anthropics/financial-services: 62 skills, 10 agents and 27 commands as 99 skills; the MCP connectors an agent needs are named in the skill, not installed). Pick a whole pack or single skills; one owner per skill name still holds.
+- Design.md installs at most 3 skills at a time. A 4th pick stops with a choice: `--drop <id>` (repeatable), or in the wizard a list where Enter drops one and Esc keeps all. `--preset design-md` is refused and `--preset all` skips it.
+- Pack sources are fetched in parallel (up to four at once), sparse and shallow, and a pinned source is cached in `~/.sigmaskills/cache/git`. A pack source can set `transform` and a pack can set `limit`.
+- Wizard: key `0` picks the tenth pack; a limited pack shows `held/3`; installed skills are marked; the filter matches a skill's group.
+- `list --packs` stops long lists after eight rows and notes a pack's limit. `list --packs --json` adds `limit` and `group`.
+- `scripts/packs-snapshot.js` keeps the packs it does not list.
+
 ## [0.16.7] — 2026-10-23
 
 ### Fixed
@@ -490,7 +544,9 @@ Ship high-rigor, installable agent skills (not apps): full-repo audits, performa
 | 2026-08-11 | `5dab832` | Add CHANGELOG and mark v0.1.0 as the first Sigmaskills release |
 | 2026-08-11 | `124a666` | Polish v0.1.0 docs for multi-host install and use |
 
-[Unreleased]: https://github.com/Djordje-Stojanovic/Sigmaskills/compare/v0.16.7...HEAD
+[Unreleased]: https://github.com/Djordje-Stojanovic/Sigmaskills/compare/v0.16.9...HEAD
+[0.16.9]: https://github.com/Djordje-Stojanovic/Sigmaskills/compare/v0.16.8...v0.16.9
+[0.16.8]: https://github.com/Djordje-Stojanovic/Sigmaskills/compare/v0.16.7...v0.16.8
 [0.16.7]: https://github.com/Djordje-Stojanovic/Sigmaskills/compare/v0.16.6...v0.16.7
 [0.16.6]: https://github.com/Djordje-Stojanovic/Sigmaskills/compare/v0.16.5...v0.16.6
 [0.16.5]: https://github.com/Djordje-Stojanovic/Sigmaskills/compare/v0.16.4...v0.16.5

package/README.md

@@ -6,7 +6,7 @@ Portable [Agent Skills](https://agentskills.io/) that install once and run on **
 
 | | |
 |---|---|
-| **Release** | [**v0.16.7**](https://github.com/Djordje-Stojanovic/Sigmaskills/releases/tag/v0.16.7) |
+| **Release** | [**v0.16.9**](https://github.com/Djordje-Stojanovic/Sigmaskills/releases/tag/v0.16.9) |
 | **Changelog** | [CHANGELOG.md](CHANGELOG.md) |
 | **License** | [MIT](LICENSE) |
 | **Spec** | [agentskills.io](https://agentskills.io/) |
@@ -22,6 +22,14 @@ Portable [Agent Skills](https://agentskills.io/) that install once and run on **
      arena      →  blind pairs that rank any options by Elo
   Σ  research   →  read public sources, record evidence and access gaps
   Σ  deeplook   →  helper for what agents cannot read: video, speech, scans, binaries
+  Σ  loop       →  an unattended-run prompt and a loop file from open work
+  Σ  taste      →  one ranked plan of big visible changes from 14 lenses
+  Σ  human      →  text that reads like a person wrote it, every fact kept
+  Σ  fork       →  an isolated copy of any repo and a course on it
+  Σ  design     →  a design system from one accent colour
+  Σ  tour       →  a visual tour of any repo in one page
+  Σ  financeman →  DCF, LBO, comps and budgets from SEC filings
+  Σ  freedom    →  deletes model-written refusals from a repo's agent docs
 ```
 
 ---
@@ -49,6 +57,10 @@ The wizard has three steps: **What**, **Where**, **Review**. It writes nothing u
 | `threejs` | 13 Three.js skills | [CloudAI-X/threejs-skills](https://github.com/CloudAI-X/threejs-skills) |
 | `hyperframes` | HyperFrames and the student kit | [heygen-com/hyperframes](https://github.com/heygen-com/hyperframes), [nateherkai/hyperframes-student-kit](https://github.com/nateherkai/hyperframes-student-kit) |
 | `minimax-h3` | MiniMax H3 prompt and recipe skills | [MiniMax-AI/MiniMax-H3](https://github.com/MiniMax-AI/MiniMax-H3) |
+| `anthropic` | 19 Anthropic skills: docx, pdf, pptx, xlsx, Claude API, MCP builder, skill creator | [anthropics/skills](https://github.com/anthropics/skills) |
+| `design-md` | 84 brand looks, each a `DESIGN.md` skill. **Up to 3 at a time.** | [VoltAgent/awesome-design-md](https://github.com/VoltAgent/awesome-design-md) |
+| `last30days` | research any topic across the last 30 days (needs Python 3.12+ and Node 22+) | [mvanhorn/last30days-skill](https://github.com/mvanhorn/last30days-skill) |
+| `financial-services` | 99 finance skills: DCF, LBO, comps, pitch decks, earnings, KYC (plugins, agents and commands as skills) | [anthropics/financial-services](https://github.com/anthropics/financial-services) |
 | `local` | a folder of your own skills | `SIGMA_LOCAL_SKILLS_DIR` or `~/.sigmaskills/packs.json` |
 | `all` | every pack that is available | |
 
@@ -61,6 +73,8 @@ npx @djordje-stojanovic/sigmaskills install sigmawrite --project .
 npx @djordje-stojanovic/sigmaskills install --preset sigma --preset mattpocock --agent claude --project .
 npx @djordje-stojanovic/sigmaskills install --preset all --global --yes
 npx @djordje-stojanovic/sigmaskills list --packs
+npx @djordje-stojanovic/sigmaskills install design-stripe design-linear-app --project .
+npx @djordje-stojanovic/sigmaskills install design-vercel --drop design-stripe --project .
 npx @djordje-stojanovic/sigmaskills update --global --yes
 npx @djordje-stojanovic/sigmaskills uninstall --skill tdd --yes
 ```
@@ -77,6 +91,7 @@ The [installer guide](https://github.com/Djordje-Stojanovic/Sigmaskills/blob/mai
 | **SigmaImprove** | `sigmaimprove` | Find what would make it clearly better, measured against the world's best | One living issue per category (1–20), updated on every run |
 | **SigmaBrief** | `sigmabrief` | Prompt factory for parallel / single agents | Chat briefs only |
 | **SigmaShip** | `sigmaship` | Ship one planned ticket: build, review rounds, merge, clean up | One merged PR, closed ticket, main in sync |
+| **SigmaFreedom** | `sigmafreedom` | De-safeguard a repo: delete model-written refusals and moral limits from agent docs, keep a lint so they never return | A was/now change log and a keep-list for the owner's veto |
 | **SigmaWrite** | `sigmawrite` | Clear STE-inspired technical English | Chat writing voice |
 | **SigmaRefactor** | `sigmarefactor` | Safe, user-approved large-file refactoring | Discussion, changes, and verification |
 | **SigmaWriteGUI** | `sigmawritegui` | Plain chat with status and question boards | Interactive HTML board and copied replies |
@@ -266,6 +281,7 @@ $sigmafinanceman DCF and comps for MSFT against AAPL, GOOGL and META
 | `sigmaimprove` | repo · path · running product | One setup message (vision, focus, constraints, ticket place), then discusses ideas |
 | `sigmabrief` | issue URL · `#N` · `all open` · plain work | Only when you ask for it |
 | `sigmaship` | issue URL · `#N` | Refuses tickets without checkable acceptance criteria |
+| `sigmafreedom` | a repo path | Deletes refusals from agent docs and skill sources; every deletion is in a was/now log |
 | `sigmawrite` | (no input) | Session writing voice until you turn it off |
 | `sigmarefactor` | repository path or current repository | Scans and discusses large files before any edit |
 | `sigmawritegui` | “easy talk” · status or question batch · `EASYTALK REPLY` | Conversation mode until “easy talk off” |
@@ -299,6 +315,10 @@ $sigmafinanceman DCF and comps for MSFT against AAPL, GOOGL and META
 | SigmaFork | Python 3 · `git` · `gh` for a fork · Node.js |
 | SigmaTaste | Read access to the target · a browser or screenshot tool to see a page; without one the plan lists what was not judged |
 | SigmaHuman | Nothing beyond chat · Python 3 (standard library) for the optional fact check |
+| SigmaDesign | Python 3 (standard library) · a browser to look at the sample page |
+| SigmaTour | Python 3 (standard library) · `git` · a browser for the page |
+| SigmaFinanceMan | Python 3.10 or newer with `openpyxl` and `python-pptx` · LibreOffice is optional (`audit`, `render`) |
+| SigmaFreedom | Python 3 (standard library) for `guardrail_lint.py` |
 | SigmaDeepLook | Python 3.10 or newer; optional ffmpeg, LibreOffice, poppler, tesseract and Python packages · `look.py doctor` lists them; missing tools become Not-looked-at rows |
 
 Skills use the shell native to the machine that runs the agent (PowerShell on Windows, POSIX shell elsewhere). They do not assume WSL.

package/manifest.json

@@ -1,7 +1,7 @@
 {
   "schemaVersion": 1,
   "name": "sigmaskills",
-  "version": "0.16.7",
+  "version": "0.16.9",
   "description": "Portable Agent Skills monorepo and first-party Sigma Installer",
   "skills": [
     {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@djordje-stojanovic/sigmaskills",
-  "version": "0.16.7",
+  "version": "0.16.9",
   "description": "Portable Agent Skills monorepo and first-party Sigma Installer",
   "type": "module",
   "license": "MIT",