claude-code-telegrammer 0.4.2__tar.gz → 0.5.3__tar.gz

claude_code_telegrammer-0.5.3/.claude/commands/.old/audit_v01.md

@@ -0,0 +1,161 @@
+## Audit
+
+Is everything clean and professional? For example:
+
+- [ ] Pyproject.toml
+  - [ ] This is the place where first author is credited. For me, use:
+    - [ ] Yusuke Watanabe
+    - [ ] ywatanabe@scitex.ai
+
+- [ ] Documents
+  - [ ] Organized, up-to-date, not redundant, necessary and sufficient
+
+- [ ] Shell Script
+  - [ ] Have argparser, usage command, help option
+
+- [ ] Python API
+  - [ ] Internal code are not exposed to users, minimizing APIs for better user experience
+
+- [ ] CLI commands
+  - [ ] No original logics - always use Python or shell logics as is
+  - [ ] -h | --help option must be available all for all commands
+  - [ ] --help-recursive option must be available for all commands with children
+  - [ ] Ensure standardized naming
+  - [ ] Intuitive, organized
+  - [ ] Cli command is equipped with tab completion
+    
+- [ ] HTTP Service API
+  - [ ] No original logics - always delegate to CLI commands
+
+- [ ] MCP Service API
+  - [ ] No original logics - always delegate to CLI commands
+  - [ ] What AI agent called must be always reproducible by humans with the corresponding CLI command
+  - [ ] Standardized sub-commands:
+    $ package-name mcp {start,doctor,installation,list-tools}
+    See ~/proj/scitex-code
+  - [ ] For MCP server docs (readme, readthedocs), env var/src examples and setup please learn from ~/proj/scitex-audio
+    - [ ] /home/ywatanabe/proj/scitex-audio/GITIGNORED/LESSONS.md
+
+- [ ] Tests
+  - [ ] Is coverage calculated?
+  - [ ] Is coverage sufficient?
+
+- [ ] CI
+  - [ ] Is CI correctly setup?
+  - [ ] Is the last CI passed? If failed, are they already fixed?
+
+- [ ] Reproducible without developer's memory
+  - [ ] See `/no-long-term-memory` command
+
+- [ ] Cleanliness
+  - [ ] Is project root clean without unnecessary artifacts?
+
+- [ ] The project will work as expected and documented
+  - [ ] Run small experiments for verification if needed 
+
+- [ ] Version consistency
+  - [ ] toml, __init__.py, tag, release, pypi and so on
+
+- [ ] PyPI
+  - [ ] .github/workflows/publish-pypi.yml
+    - [ ] environment pypi
+    - [ ] First you need to publish to pypi manually. This is required to configure trusted publisher at pypi.org; project itself is not recognized, displayed otherwise.
+    - [ ] So, first check if pypi is serving the package and if not, request to the user for configuration
+
+- [ ] No personal info
+  - [ ] If package is designed for publication, do not include my own setups and keep generic tones
+  - [ ] .env contents (gitignored), name, email, github should be accepted
+  
+- [ ] Examples
+  - [ ] `./examples` must have demonstrations for main features with numbering like:
+    - [ ] `./examples/00_run_all.sh`
+    - [ ] `./examples/01_<descriptive-name>.{py,ipynb,sh}`
+  - [ ] Artifacts must be saved close place
+    - [ ] `./examples/01_<descriptive-name>_out/`
+    - [ ] Artifacts should be also included in git (GitHub)
+
+- [ ] Environmental Variables
+  - [ ] Safe for name conflict with prefix (e.g., NG: "ENV_NAME", OK: "PROJECT_NAME_ENV_NAME")
+  - [ ] PROJECT_NAME_DEBUG_MDOE=1
+  - [ ] .env file in project root
+
+- [ ] GitHub About Section
+  - [ ] Description, Homepage, and Topics are well-written for the current codebase
+  - [ ] Match user expectations and actual implementations
+  - [ ] Consider SEO effectiveness as well
+    - [ ] Add `scitex` to keywords for scitex ecosystem package
+
+- [ ] SciTeX Brand
+  - [ ] Keep consistency in cli commands
+  - [ ] Use fastapi and fastmcp when needed
+  - [ ] For delegation please check branding changer logics
+  - [ ] Do not add ywatanabe@scitex.ai "on footer" of readme
+    - [ ] This is new rule to show scitex is not for my project but for the community
+
+  ## Delegation to Downstream Packages
+  django (~/proj/scitex-cloud) -> scitex (~/proj/scitex-python) -> downstream packages like:
+  ~/proj/figrecipe (= scitex-plt)
+  ~/proj/scitex-writer
+  ~/proj/scitex-dataset
+  ~/proj/socialia
+  ~/proj/scitex-stats
+  ~/proj/scitex-clew
+  ~/proj/scitex-io
+  ...
+  ~/proj/scitex-dev (this should have all packages listed in the source code to provide ecosystem-wide operations)
+
+  Business logics to the most downstream module and upper modules should delegate to downstream packages like a cascade. Keep separation of concerns and do not repeat yourself for single source of truth.
+
+## SciTeX Standalone, Downstream Packages
+- [ ] Confirm the original, orchestrate scitex package (~/proj/scitex-python) works as before
+- [ ] Ensure always worked local develop -> origin/develop -> origin/main
+  - [ ] pip version starts from v0.1.0
+- [ ] SciTeX packages must have python apis (minimal), cli commands, and mcp tools:
+  - [ ] $ scitex-xxx --help-recursive (in any depth)
+  - [ ] $ scitex-xxx list-python-apis -v|-vv|-vvv
+  - [ ] $ scitex-xxx mcp list-tools -v|-vv|-vvv
+  - [ ] MCP uses fastmcp and the main scitex package ~/proj/scitex-python correctly delegates to the standalone package without any hard coding
+  - [ ] After MCP updated, let me know to reconnect scitex mcp server
+- [ ] LICENSE is AGPL v3.0 only
+  - [ ] CLA.md and CONTRIBUTING.md placed as well like scitex-python
+- [ ] Ensure remote is public
+- [ ] Read the Docs implemented correctly
+  - [ ] Use `$READTHEDOCS_TOKEN` for API access to RTD
+  - [ ] First, push to GitHub with the RTD conifg file and source. 
+  - [ ] Then, please let me know. I will configure to trigger RTD build on PR to main
+  - [ ] If you can register trusted publisher, please set by yourself
+- [ ] Before pushing to origin/develop, please ensure "/audit" passed
+- [ ] Ensure README.md follows the same format with https://github.com/ywatanabe1989/scitex-clew/README.md
+  - [ ] Ensure that README.md organization is as follows. 
+        Note: Add custom sections flexibly
+        Note: Use <details>, <summary>, tables, figures effectively
+        Note: Figure and Tables need legend like in scientific papers
+        Note: Figures must be visible in both dark/light modes
+        Note: Ensure icons are correctly git-tracked and pushed
+        Note: Acronyms must be spelled out on their first appearance.
+    - [ ] Project Title (e.g., `SciTeX Stats`)
+    - [ ] Header - Logo, Description, badges, RTD Link, pip install xxx
+    - [ ] Problem (scope definition, problem description)
+    - [ ] Solution (How <package-name> solves the problem)
+    - [ ] No `import scitex as stx`; use `import scitex` instead in READMEs and RTD
+    - [ ] Installation
+    - [ ] Quick Start
+    - [ ] Three Interfaces
+      - [ ] Python API
+      - [ ] CLI Commands
+      - [ ] MCP Server
+    - [ ] Part of SciTeX
+      - [ ] <package-name> is part of SciTeX. When used in side the orchestrator package `scitex`, synergy between modules can be enjoyed: (revise this sentence based on the package and synergy packages with example code)
+      - [ ] If synergy is not expected skip the "When used in ..." section., skip this as the Python API section above will suffice.
+      - [ ] The SciTeX ecosystem follows the Four Freedoms for researchers, inspired by [the Free Software Definition](https://www.gnu.org/philosophy/free-sw.en.html):
+        - [ ] ...
+    - [ ] Footer with scitex icon
+- [ ] ./examples implemented
+- [ ] ./tests implemented
+- [ ] Add `scitex_dev` as dependency in each package's `pyproject.toml`
+- [ ] Use `@supports_return_as` decorator:
+- [ ] Unified MCP response format across all mounted MCP servers
+- [ ] Wire `docs` entry points in all packages
+
+
+If you find room for improvement, do not hesitate but keep on working the remaining tasks

claude_code_telegrammer-0.5.3/.claude/commands/.old/audit_v02-numbering.md

@@ -0,0 +1,178 @@
+## Audit
+
+Is everything clean and professional? For example:
+
+- [ ] Pyproject.toml
+  - [ ] This is the place where first author is credited. For me, use:
+    - [ ] Yusuke Watanabe
+    - [ ] ywatanabe@scitex.ai
+
+- [ ] Documents
+  - [ ] Organized, up-to-date, not redundant, necessary and sufficient
+
+- [ ] Shell Script
+  - [ ] Have argparser, usage command, help option
+
+- [ ] Python API
+  - [ ] Internal code are not exposed to users, minimizing APIs for better user experience
+
+- [ ] CLI commands
+  - [ ] No original logics - always use Python or shell logics as is
+  - [ ] -h | --help option must be available all for all commands
+  - [ ] --help-recursive option must be available for all commands with children
+  - [ ] Ensure standardized naming
+  - [ ] Intuitive, organized
+  - [ ] Cli command is equipped with tab completion
+    
+- [ ] HTTP Service API
+  - [ ] No original logics - always delegate to CLI commands
+
+- [ ] MCP Service API
+  - [ ] No original logics - always delegate to CLI commands
+  - [ ] What AI agent called must be always reproducible by humans with the corresponding CLI command
+  - [ ] Standardized sub-commands:
+    $ package-name mcp {start,doctor,installation,list-tools}
+    See ~/proj/scitex-code
+  - [ ] For MCP server docs (readme, readthedocs), env var/src examples and setup please learn from ~/proj/scitex-audio
+    - [ ] /home/ywatanabe/proj/scitex-audio/GITIGNORED/LESSONS.md
+
+- [ ] Tests
+  - [ ] Is coverage calculated?
+  - [ ] Is coverage sufficient?
+
+- [ ] CI
+  - [ ] Is CI correctly setup?
+  - [ ] Is the last CI passed? If failed, are they already fixed?
+
+- [ ] Reproducible without developer's memory
+  - [ ] See `/no-long-term-memory` command
+
+- [ ] Cleanliness
+  - [ ] Is project root clean without unnecessary artifacts?
+
+- [ ] The project will work as expected and documented
+  - [ ] Run small experiments for verification if needed 
+
+- [ ] Version consistency
+  - [ ] toml, __init__.py, tag, release, pypi and so on
+
+- [ ] PyPI
+  - [ ] .github/workflows/publish-pypi.yml
+    - [ ] environment pypi
+    - [ ] First you need to publish to pypi manually. This is required to configure trusted publisher at pypi.org; project itself is not recognized, displayed otherwise.
+    - [ ] So, first check if pypi is serving the package and if not, request to the user for configuration
+
+- [ ] No personal info
+  - [ ] If package is designed for publication, do not include my own setups and keep generic tones
+  - [ ] .env contents (gitignored), name, email, github should be accepted
+  
+- [ ] Examples
+  - [ ] `./examples` must have demonstrations for main features with numbering like:
+    - [ ] `./examples/00_run_all.sh`
+    - [ ] `./examples/01_<descriptive-name>.{py,ipynb,sh}`
+  - [ ] Artifacts must be saved close place
+    - [ ] `./examples/01_<descriptive-name>_out/`
+    - [ ] Artifacts should be also included in git (GitHub)
+
+- [ ] Environmental Variables
+  - [ ] Safe for name conflict with prefix (e.g., NG: "ENV_NAME", OK: "PROJECT_NAME_ENV_NAME")
+  - [ ] PROJECT_NAME_DEBUG_MDOE=1
+  - [ ] .env file in project root
+
+- [ ] GitHub About Section
+  - [ ] Description, Homepage, and Topics are well-written for the current codebase
+  - [ ] Match user expectations and actual implementations
+  - [ ] Consider SEO effectiveness as well
+    - [ ] Add `scitex` to keywords for scitex ecosystem package
+
+- [ ] SciTeX Brand
+  - [ ] Keep consistency in cli commands
+  - [ ] Use fastapi and fastmcp when needed
+  - [ ] For delegation please check branding changer logics
+  - [ ] Do not add ywatanabe@scitex.ai "on footer" of readme
+    - [ ] This is new rule to show scitex is not for my project but for the community
+
+  ## SciTeX Package Architecture (3-Layer Cascade)
+
+  See full rules: `~/proj/scitex-dev/docs/MASTER/00_SCITEX_UPSTREAM_AND_DOWNSTREAM_RULES.md`
+
+  ```
+  Downstream (apps — standalone, own IO):
+    ~/proj/figrecipe, ~/proj/scitex-writer, ~/proj/scitex-clew, ...
+        ↑ wraps/cascades via plugin registry
+  Middle (shared infrastructure):
+    ~/proj/scitex-io, ~/proj/scitex-app, ~/proj/scitex-ui,
+    ~/proj/scitex-stats, ~/proj/scitex-audio, ~/proj/scitex-dev
+        ↑ integrates/re-exposes (SOC — integration tests ONLY)
+  Upstream (orchestration):
+    ~/proj/scitex-python (scitex), ~/proj/scitex-cloud
+  ```
+
+  ### Key Rules
+  - [ ] **Apps work standalone** — no scitex dependency for core functionality
+  - [ ] **scitex-io wraps, not replaces** — cascades through plugin registry
+  - [ ] **scitex re-exposes only** — no logic, just re-export from middle layer
+  - [ ] **Upstream has ONLY integration tests** — unit tests belong downstream
+  - [ ] **All 3 interfaces cascade** — Python API, CLI, MCP server same direction
+  - [ ] **_AVAILABLE flags** — detect optional deps, provide install instructions
+  - [ ] **extras in pyproject.toml** — `pip install figrecipe[scitex]`
+  - [ ] **Never reverse imports** — upstream never imports downstream directly
+
+  ### Testing Scope
+  - [ ] Downstream (figrecipe): unit tests for own logic
+  - [ ] Middle (scitex-io): integration tests for cascade/wrapping
+  - [ ] Upstream (scitex): integration tests ONLY for full pipeline
+
+## SciTeX Standalone, Downstream Packages
+- [ ] Confirm the original, orchestrate scitex package (~/proj/scitex-python) works as before
+- [ ] Ensure always worked local develop -> origin/develop -> origin/main
+  - [ ] pip version starts from v0.1.0
+- [ ] SciTeX packages must have python apis (minimal), cli commands, and mcp tools:
+  - [ ] $ scitex-xxx --help-recursive (in any depth)
+  - [ ] $ scitex-xxx list-python-apis -v|-vv|-vvv
+  - [ ] $ scitex-xxx mcp list-tools -v|-vv|-vvv
+  - [ ] MCP uses fastmcp and the main scitex package ~/proj/scitex-python correctly delegates to the standalone package without any hard coding
+  - [ ] After MCP updated, let me know to reconnect scitex mcp server
+- [ ] LICENSE is AGPL v3.0 only
+  - [ ] CLA.md and CONTRIBUTING.md placed as well like scitex-python
+- [ ] Ensure remote is public
+- [ ] Read the Docs implemented correctly
+  - [ ] Use `$READTHEDOCS_TOKEN` for API access to RTD
+  - [ ] First, push to GitHub with the RTD conifg file and source. 
+  - [ ] Then, please let me know. I will configure to trigger RTD build on PR to main
+  - [ ] If you can register trusted publisher, please set by yourself
+- [ ] Before pushing to origin/develop, please ensure "/audit" passed
+- [ ] Ensure README.md follows the same format with https://github.com/ywatanabe1989/scitex-clew/README.md
+  - [ ] Ensure that README.md organization is as follows. 
+        Note: Add custom sections flexibly
+        Note: Use <details>, <summary>, tables, figures effectively
+        Note: Figure and Tables need legend like in scientific papers
+        Note: Figures must be visible in both dark/light modes
+        Note: Ensure icons are correctly git-tracked and pushed
+        Note: Acronyms must be spelled out on their first appearance.
+    - [ ] Project Title (e.g., `SciTeX Stats`)
+    - [ ] Header - Logo, Description, badges, RTD Link, pip install xxx
+    - [ ] Problem (scope definition, problem description)
+    - [ ] Solution (How <package-name> solves the problem)
+    - [ ] No `import scitex as stx`; use `import scitex` instead in READMEs and RTD
+    - [ ] Installation
+    - [ ] Quick Start
+    - [ ] Three Interfaces
+      - [ ] Python API
+      - [ ] CLI Commands
+      - [ ] MCP Server
+    - [ ] Part of SciTeX
+      - [ ] <package-name> is part of SciTeX. When used in side the orchestrator package `scitex`, synergy between modules can be enjoyed: (revise this sentence based on the package and synergy packages with example code)
+      - [ ] If synergy is not expected skip the "When used in ..." section., skip this as the Python API section above will suffice.
+      - [ ] The SciTeX ecosystem follows the Four Freedoms for researchers, inspired by [the Free Software Definition](https://www.gnu.org/philosophy/free-sw.en.html):
+        - [ ] ...
+    - [ ] Footer with scitex icon
+- [ ] ./examples implemented
+- [ ] ./tests implemented
+- [ ] Add `scitex_dev` as dependency in each package's `pyproject.toml`
+- [ ] Use `@supports_return_as` decorator:
+- [ ] Unified MCP response format across all mounted MCP servers
+- [ ] Wire `docs` entry points in all packages
+
+
+If you find room for improvement, do not hesitate but keep on working the remaining tasks

claude_code_telegrammer-0.5.3/.claude/commands/actions.md

@@ -0,0 +1,28 @@
+# Check GitHub actions and suggest how we can improve the quality of this project
+```bash
+gh repo set-default <github-username>/<repository-name>
+# e.g., gh repo set-default ywatanabe1989/scitex-code
+
+# List recent workflow runs
+gh run list --limit 10 --branch develop
+
+#View spefic run details
+gh run view <RUN_ID>
+# e.g., gh run view 18892191254
+
+# Get logs from a run
+gh run view <RUN_ID> --log
+# e.g., gh run view 18892191254 --log
+
+# Get specific workflow information
+gh workflow list
+gh workflow view <WORKFLOW_ID>
+```
+
+# Check Read the Docs Builds Status and fix if needed
+
+## Project Build Status Page
+https://app.readthedocs.org/projects/<package-name>/builds/<build-id>/
+
+## Full Log
+https://app.readthedocs.org/api/v2/build/<build-id>.txt

claude_code_telegrammer-0.5.3/.claude/commands/call.md

@@ -0,0 +1,2 @@
+#### Alarm Call
+User might be sleeping. When you speak more than 10 times but no response returned by the user, please call to the user, me (+61-461-522-907).

claude_code_telegrammer-0.5.3/.claude/commands/checkmarks.md

@@ -0,0 +1,2 @@
+Make a checklist with check marks ([ ])
+$ARGUMENTS

claude_code_telegrammer-0.5.3/.claude/commands/diagram.md

@@ -0,0 +1,7 @@
+# Diagram
+
+Create diagram based on user requests.
+Check available tools in this order of priority.
+
+1. `$ scitex plt --version`
+2. `$ figrecipe --version`

claude_code_telegrammer-0.5.3/.claude/commands/migrate-to-skills.md

@@ -0,0 +1,2 @@
+If commands are more suitable in skills, please migrate them:
+$ARGUMENTS

claude_code_telegrammer-0.5.3/.claude/commands/model_update.md

@@ -0,0 +1,6 @@
+Recently, new models announced from LLM providers ($ARGUMENTS).
+
+Please add the new models and update their costs in:
+
+- scitex.ai.GenAI (`~/proj/scitex_repo/src/scitex/ai/_gen_ai/_PARAMS.py`)
+- genai (`~/.emacs.d/lisp/genai/`)

claude_code_telegrammer-0.5.3/.claude/commands/no-fallbacks.md

@@ -0,0 +1,11 @@
+## No fallbacks and No silent failures
+- [ ] Do never allow for silent failures.
+- [ ] Do never implement fallbacks unless user explicitly requests
+- [ ] This is crucial to show errors as errros without hiding actual problems.
+  - [ ] For fallbacks user confirmed, please comment it in source code.
+  - [ ] Otherwise, no fallbacks accepted.
+- [ ] Red cases are useful. We have opportunities to improve. Do not hide them.
+- [ ] False positives are really difficult to find and fix. Do not workaround problems silently.
+- [ ] Not working must be not working.
+- [ ] Not complete must be not complete.
+- [ ] Just make everything clear and honest without false information.

claude_code_telegrammer-0.5.3/.claude/commands/read-the-docs.md

@@ -0,0 +1,21 @@
+# Check Read the Docs Builds Status and fix if needed
+
+## Project Build Status Page
+https://app.readthedocs.org/projects/<package-name>/builds/<build-id>/
+
+## How to Trigger Builds
+PR to origin/main
+
+## Full Log (this is long but useful for identifying actual error)
+https://app.readthedocs.org/api/v2/build/<build-id>.txt
+
+## Note
+- Follow the format of already accomplished projects
+  ~/proj/figrecipe
+  ~/proj/scitex-writer
+- Especially, install[all] is the key. It would be heavy but no package error is much better.
+- Often encountered issue:
+  ```
+  python -m pip install --exists-action=w --no-cache-dir -r docs/sphinx/requirements.txt  
+  ERROR: Could not open requirements file: [Errno 2] No such file or directory: 'docs/sphinx/requirements.txt'
+  ```

claude_code_telegrammer-0.5.3/.claude/commands/s.md

@@ -0,0 +1 @@
+Load and apply the skill(s): @ARGUMENTS

claude_code_telegrammer-0.5.3/.claude/commands/speak-and-call.md

@@ -0,0 +1,28 @@
+## Skills to Load (Required)
+
+Load these skills first:
+
+```skills
+quality-guards
+autonomous
+scitex-audio
+scitex-notification
+speak-and-signature
+```
+
+## Override:
+
+- Continue working on the task.
+  - Until requirements have satisfied and confirmed visually and functionally, 
+
+- No false positive reports.
+
+- Speak in English
+
+- Make a phone call to the user if you speak seven times but the user does not respond
+
+
+## Requirements are below:
+$ARGUMENTS
+
+(If no requirements shown here, please estimate requests based on the tasks so far)

claude_code_telegrammer-0.5.3/.claude/commands/tmp-txt.md

@@ -0,0 +1 @@
+Read ~/win/downloads/tmp.txt

claude_code_telegrammer-0.5.3/.claude/commands/todo.md

@@ -0,0 +1 @@
+Write TODO to ~/scitex-todo based on the recent discussion, 

claude_code_telegrammer-0.5.3/.claude/commands/tree.md

@@ -0,0 +1 @@
+Check the repository structure by running `tree --gitignore`

claude_code_telegrammer-0.5.3/.claude/commands/tweet.md

@@ -0,0 +1,3 @@
+Using the `scitex mcp server`, tweet about the following package/project/ideas with language/content revision. No need for my confirmation, just post your revised version(s). Threading is okay. Follow effective patterns for platforms.
+
+$ARGUMENTS

claude_code_telegrammer-0.5.3/.claude/commands/understand-guidelines.md

@@ -0,0 +1,26 @@
+0. Check the current worktree is appropriate:
+   Follow the version control guideline:
+   `./docs/to_claude/guidelines/**/*version*control*.md`
+
+Once worktree is confirmed as valid, proceed to the following steps:
+1. Understand the project structure
+   `tree --gitignore ./{src,scripts,tests,examples,project_management}`
+2. Always try to use the `filesystem` mcp as it is faster than `Read`
+   e.g., `filesystem:read_multiplefiles`
+3. Read and understand relevant guidelines, scripts, and so on:
+   e.g., `./docs/to_claude/guidelines/**/*.md`
+   e.g., `./docs/to_claude/bin/**`
+   e.g., ANY FILES YOU THINK IMPORTANT
+5. Summarize what you learned from these steps.
+6. Memorize the titles of these guidelines and scripts to recall in anytime.
+
+# Final check
+7. List all files you read in this request:
+``` plaintext
+Read and understand: `/path/to/read/file1`
+Read and understand: `/path/to/read/file2`
+...
+
+```
+
+8. Summarize what you understood using bullet points

claude_code_telegrammer-0.5.3/.claude/commands/update-docs.md

@@ -0,0 +1,10 @@
+Based on recent changes, update documents:
+
+## Locations
+
+- README.md (potentially in various levels)
+- Read the Docs
+- Other related documentations
+- src/scripts/examples/tests
+
+$ARGUMENTS

claude_code_telegrammer-0.5.3/.claude/commands/update-scitex-full.md

@@ -0,0 +1,24 @@
+## Skills to Load (Required)
+skill:scitex-dev-full-update
+skill:speak-and-signature
+skill:autonomous
+skill:quality-guards
+
+## Task: Full Ecosystem Release
+
+Follow the full-update skill strictly — 5 phases, 20 steps, in order.
+
+1. **Phase 1: Pre-flight** (steps 1–4) — check_ci, detect_mismatches, fix_init_version, verify_pypi_config
+2. **Phase 2: Version Bump + Release** (steps 5–12) — determine_bump_type, bump_version, verify_docs_and_skills, commit_and_tag, check_ci_before_push, push, create_github_release, wait_for_pypi
+3. **Phase 3: Local Sync** (steps 13–14) — fix_local, verify_versions
+4. **Phase 4: Host Sync** (steps 15–17) — fix_remote [Custom], deploy_scitex_cloud [Custom], verify_production [Custom]
+5. **Phase 5: Verification** (steps 18–20) — verify_versions, verify_production [Custom], report_summary
+
+Each step has Python API, CLI, and MCP tool — use the most appropriate for the context.
+Use parallel subagents for independent packages within a phase.
+Present ecosystem summary table when done.
+
+## Arguments
+$ARGUMENTS
+
+(If no arguments, process all packages with appropriate version increments)

claude_code_telegrammer-0.5.3/.claude/commands/update-scitex.md

@@ -0,0 +1,14 @@
+## Skills to Load (Required)
+skill:scitex-versions
+skill:speak-and-signature
+
+## Task
+1. Check the levels of version management based on the skill.
+2. Investigate current states for this project.
+3. Present which level is the best match at the moment.
+4. List the numbered choices and wait for user selection.
+
+## Arguments
+$ARGUMENTS
+
+(If no arguments, run `scitex dev versions list` and report current state)

claude_code_telegrammer-0.5.3/.claude/commands/update-skills.md

@@ -0,0 +1,36 @@
+Based on recent experiences, create or update skills.
+
+## Locations
+
+### General (non-SciTeX):
+- ~/.dotfiles/src/.claude/to_claude/skills
+
+### SciTeX Modules (submodules within scitex-python):
+- ~/proj/scitex-python/src/scitex/<module-name>/_skills
+
+### SciTeX Standalone Packages:
+- ~/proj/<package-name>/src/<package_name>/_skills/<pip-name>/
+
+## SciTeX-Specific Rules
+
+- SKILL.md is an index only — content goes in focused sub-skill files
+- Include MANIFEST.md with version (auto-stamped from pyproject.toml on export)
+- Each SKILL.md should reference MANIFEST.md for update instructions
+- Use `<details>/<summary>` in READMEs for "Four Interfaces" (not "Three")
+
+### After Updating SciTeX Skills
+
+```bash
+scitex-dev skills list --package <pkg>
+scitex-dev skills export --package <pkg> --dry-run
+python -m pytest tests/test_skills.py -v
+```
+
+## General Skill-Writing Rules
+
+- No checkboxes (`- [ ]`) — use prose, code blocks, tables
+- No "Confirmation" / "Understanding Check" sections
+- Show `$ENV_VAR_NAME` in CLI help defaults, not resolved values
+- Consolidate thin sub-files (< 20 lines each) into SKILL.md
+
+$ARGUMENTS

claude_code_telegrammer-0.5.3/.claude/hooks/.gitignore

@@ -0,0 +1,2 @@
+.claude
+docs/to_claude