@hasna/assistants 1.1.99 → 1.1.101

package/package.json

@@ -1,6 +1,7 @@
 {
   "name": "@hasna/assistants",
-  "version": "1.1.99",
+
+  "version": "1.1.101",
   "description": "Personal AI assistant that runs in your terminal — powered by Claude",
   "type": "module",
   "bin": {

package/dist/.assistants/ASSISTANTS.md

@@ -1,7 +0,0 @@
-You are an AI assistant running in the terminal.
-
-Rules:
-- Use connector tools for external systems when available.
-- Prefer read-only tools (read, glob, grep) for inspection.
-- Only write helper scripts in `.assistants/scripts/<session-id>/`.
-- Avoid destructive commands, installs, or environment changes unless explicitly asked.

package/dist/.assistants/commands/reflect.md

@@ -1,13 +0,0 @@
----
-name: reflect
-description: Reflect on the conversation and suggest next steps
-tags: [reflection, next-steps]
----
-
-# Reflection
-
-Please summarize the last interaction and suggest 2-3 next steps.
-
-- Keep it concise
-- Focus on clarity
-- Ask a follow-up question if needed

package/dist/.assistants/skills/calendar/SKILL.md

@@ -1,40 +0,0 @@
----
-name: calendar
-description: View and manage Google Calendar events
-argument-hint: [today|tomorrow|week|add "event"]
-allowed-tools: bash, googlecalendar
----
-
-## Instructions
-
-Manage calendar events based on the command: $ARGUMENTS
-
-### Commands
-
-**View today's events:**
-```bash
-connect-googlecalendar events list --today
-```
-
-**View tomorrow's events:**
-```bash
-connect-googlecalendar events list --tomorrow
-```
-
-**View this week's events:**
-```bash
-connect-googlecalendar events list --week
-```
-
-**Add an event:**
-Parse the event details from the arguments and create the event:
-```bash
-connect-googlecalendar events create "Event Title" --start "2025-01-31T10:00:00" --end "2025-01-31T11:00:00"
-```
-
-### Output Format
-
-Present events in a clean list:
-- **Time** - Event Title (Location if available)
-
-For conflicts or busy periods, highlight them.

package/dist/.assistants/skills/email/SKILL.md

@@ -1,41 +0,0 @@
----
-name: email
-description: Check and manage Gmail inbox
-argument-hint: [inbox|unread|send "to" "subject" "body"]
-allowed-tools: bash, gmail
----
-
-## Instructions
-
-Manage email based on the command: $ARGUMENTS
-
-### Commands
-
-**View inbox:**
-```bash
-connect-gmail messages list --max 10
-```
-
-**View unread:**
-```bash
-connect-gmail messages list --unread --max 10
-```
-
-**Send email:**
-Parse the recipient, subject, and body from arguments:
-```bash
-connect-gmail messages send --to "recipient@example.com" --subject "Subject" --body "Body text"
-```
-
-**Search emails:**
-```bash
-connect-gmail messages search "query" --max 10
-```
-
-### Output Format
-
-Present emails as:
-- **From** | Subject | Date
-  Preview of body (first 100 chars)
-
-Flag important or urgent emails.

package/dist/.assistants/skills/notes/SKILL.md

@@ -1,44 +0,0 @@
----
-name: notes
-description: Create and manage notes in Notion
-argument-hint: [list|create "title"|search "query"]
-allowed-tools: bash, notion
----
-
-## Instructions
-
-Manage notes in Notion based on the command: $ARGUMENTS
-
-### Commands
-
-**List recent notes:**
-```bash
-connect-notion pages list --max 10
-```
-
-**Create a note:**
-Parse the title and optional content from arguments:
-```bash
-connect-notion pages create "PARENT_PAGE_ID" "Note Title" --content "Note content here"
-```
-
-Note: You'll need to know the parent page ID. List pages first to find it.
-
-**Search notes:**
-```bash
-connect-notion search "query" --pages
-```
-
-**Get note content:**
-```bash
-connect-notion blocks children "PAGE_ID"
-```
-
-### Output Format
-
-Present notes with:
-- Title
-- Last edited date
-- Preview of content
-
-For search results, show relevance highlights.

package/dist/.assistants/skills/search/SKILL.md

@@ -1,23 +0,0 @@
----
-name: search
-description: Search the web for information using Exa
-argument-hint: [query]
-allowed-tools: bash, exa
----
-
-## Instructions
-
-Search the web for information about: $ARGUMENTS
-
-Use the `connect-exa` CLI to perform the search:
-
-```bash
-connect-exa search "$ARGUMENTS" --max 5
-```
-
-Summarize the results in a clear, concise format with:
-1. Key findings
-2. Relevant links
-3. Any important caveats
-
-If no results are found, suggest alternative search queries.

package/dist/.assistants/skills/skill-brainstorm/SKILL.md

@@ -1,30 +0,0 @@
----
-name: brainstorm
-description: Generate ideas and explore possibilities. Use when the user needs creative thinking or problem solving.
-argument-hint: [topic or problem]
----
-
-## Instructions
-
-Brainstorm ideas for: $ARGUMENTS
-
-### Approach
-
-1. **Divergent thinking**: Generate many ideas without judgment
-2. **Different angles**: Consider the topic from multiple perspectives
-3. **Build on ideas**: Create variations and combinations
-4. **Organize**: Group related ideas together
-
-### Output Format
-
-- **Quick Ideas** (5-10): Rapid-fire suggestions
-- **Deep Dives** (2-3): More developed concepts with brief explanations
-- **Wild Cards** (1-2): Unconventional or unexpected approaches
-- **Next Steps**: How to explore or refine the best ideas
-
-### Guidelines
-
-- Quantity over quality initially
-- No idea is too simple or too ambitious
-- Consider practical and creative options
-- Build on context the user provides

package/dist/.assistants/skills/skill-codereview/SKILL.md

@@ -1,46 +0,0 @@
----
-name: codereview
-description: Review code for bugs, security issues, performance problems, and style. Accepts a file path, diff, or inline code snippet.
-argument-hint: [file-path-or-diff]
-allowed-tools: Read, Bash
----
-
-## Code Review
-
-Review the following for: $ARGUMENTS
-
-### What to look for
-
-1. **Bugs & correctness** — logic errors, off-by-one, null/undefined handling, edge cases, incorrect assumptions
-2. **Security** — injection risks (SQL, shell, XSS), hardcoded secrets, insecure defaults, improper input validation, path traversal
-3. **Performance** — N+1 queries, unnecessary re-renders, blocking operations, missing indexes, large allocations
-4. **Readability** — unclear naming, overly complex logic that could be simplified, missing or misleading comments
-5. **Error handling** — swallowed errors, missing error boundaries, unclear error messages
-6. **Type safety** — `any` types, unsafe casts, missing nullability checks
-
-### Process
-
-1. If `$ARGUMENTS` is a file path, read the file first with the Read tool
-2. If it's a glob or directory, read the relevant files
-3. If it's inline code, review it directly
-4. For each issue found, state: **location** → **severity** (critical/high/medium/low) → **problem** → **fix**
-
-### Output format
-
-```
-## Code Review: <file or description>
-
-### Critical
-- [line X] <issue> → <fix>
-
-### High
-- [line X] <issue> → <fix>
-
-### Medium / Low
-- <issue> → <fix>
-
-### Summary
-<1-2 sentence overall assessment>
-```
-
-If no issues are found, say so explicitly — don't pad with non-issues.

package/dist/.assistants/skills/skill-commitmsg/SKILL.md

@@ -1,41 +0,0 @@
----
-name: commitmsg
-description: Generate a conventional commit message from staged changes or a diff. Follows Conventional Commits spec (feat/fix/chore/refactor/docs/test/perf/ci).
-argument-hint: [optional: file paths or description]
-allowed-tools: Bash
----
-
-## Generate Commit Message
-
-$ARGUMENTS
-
-### Process
-
-1. Run `git diff --staged` to see staged changes (or `git diff HEAD` if nothing staged)
-2. If `$ARGUMENTS` mentions specific files, focus on those
-3. Analyse what changed — what was added, removed, modified, and why
-4. Generate a commit message following Conventional Commits
-
-### Conventional Commits format
-
-```
-<type>(<scope>): <short summary>
-
-<optional body: what changed and why, wrapped at 72 chars>
-
-<optional footer: BREAKING CHANGE: ..., Closes #123>
-```
-
-**Types:** `feat` (new feature), `fix` (bug fix), `perf` (performance), `refactor` (no behaviour change), `docs` (documentation only), `test` (tests only), `chore` (tooling/deps), `ci` (CI/CD), `style` (formatting)
-
-**Rules:**
-- Subject line ≤ 72 characters, imperative mood ("add X" not "added X")
-- Scope is optional but useful (e.g. `feat(auth):`, `fix(cli):`)
-- Body explains *why*, not *what* (the diff shows what)
-- Mark breaking changes with `BREAKING CHANGE:` in footer
-- Reference issues with `Closes #N` or `Fixes #N`
-
-### Output
-
-Print the ready-to-use commit message in a code block so it can be copied directly.
-If the changes span multiple unrelated concerns, suggest splitting into separate commits.

package/dist/.assistants/skills/skill-debug/SKILL.md

@@ -1,59 +0,0 @@
----
-name: debug
-description: Structured debugging session. Accepts an error message, stack trace, or description of unexpected behaviour. Works through reproduce → isolate → hypothesize → verify → fix.
-argument-hint: [error message or description]
-allowed-tools: Bash, Read, Grep
----
-
-## Debug Session
-
-**Problem:** $ARGUMENTS
-
-### Step 1 — Understand the error
-
-- Parse the error message and stack trace
-- Identify the failing file, line, and function
-- Note what was expected vs. what actually happened
-
-### Step 2 — Reproduce
-
-- Find the minimal steps or input that trigger the issue
-- Run the failing command / test to confirm the error is present
-- Note exact error output
-
-### Step 3 — Isolate
-
-- Narrow down to the smallest code path involved
-- Check recent changes (`git log --oneline -10`, `git diff HEAD~1`) — did something recently break this?
-- Search for related error messages or similar patterns in the codebase
-
-### Step 4 — Hypothesize
-
-- List 2–3 plausible root causes, ranked by likelihood
-- For each hypothesis, state what evidence would confirm or rule it out
-
-### Step 5 — Verify
-
-- Test each hypothesis by reading relevant code, adding debug output, or running targeted commands
-- Eliminate hypotheses one by one until the root cause is confirmed
-
-### Step 6 — Fix
-
-- Implement the minimal correct fix
-- Verify the fix resolves the original error
-- Check for related issues the fix might cause
-- Add a regression note if useful
-
-### Output format
-
-Work through each step out loud. Show commands run and their output. End with:
-```
-## Root cause
-<one sentence>
-
-## Fix applied
-<what was changed>
-
-## Verified by
-<how you confirmed it works>
-```

package/dist/.assistants/skills/skill-draft/SKILL.md

@@ -1,30 +0,0 @@
----
-name: draft
-description: Draft emails, messages, or documents. Use when the user needs help writing something.
-argument-hint: [type] [context]
----
-
-## Instructions
-
-Draft content based on: $ARGUMENTS
-
-### Guidelines
-
-1. **Identify the type**: Email, message, letter, document, post, etc.
-2. **Understand the context**: Who is the audience? What's the purpose?
-3. **Match the tone**: Professional, casual, formal, friendly, etc.
-4. **Be concise**: Get to the point while remaining polite
-
-### Output Format
-
-Provide the draft with:
-- The complete text ready to use
-- Any placeholders marked with [BRACKETS] for info you need
-- Optional: Alternative versions if tone/approach is ambiguous
-
-### Tips
-
-- For emails: Include a clear subject line suggestion
-- For messages: Keep it appropriately brief
-- For documents: Structure with headers if lengthy
-- Always proofread for clarity and grammar

package/dist/.assistants/skills/skill-research/SKILL.md

@@ -1,27 +0,0 @@
----
-name: research
-description: Research a topic thoroughly using web search and multiple sources. Use when the user needs comprehensive information on a topic.
-argument-hint: [topic]
-allowed-tools: web_search, web_fetch, curl
----
-
-## Instructions
-
-Research the following topic: $ARGUMENTS
-
-### Steps
-
-1. Search the web for relevant information using web_search
-2. Fetch the top 3-5 most relevant results using web_fetch
-3. Cross-reference information from multiple sources
-4. Synthesize findings into a comprehensive overview
-
-### Output Format
-
-- **Overview**: Brief introduction to the topic
-- **Key Findings**: Main discoveries from research
-- **Different Perspectives**: Any varying viewpoints found
-- **Sources**: List the URLs used with brief descriptions
-- **Recommendations**: Suggested next steps or additional resources
-
-Be thorough but focused. Prioritize accuracy and cite sources.

package/dist/.assistants/skills/skill-summarize/SKILL.md

@@ -1,25 +0,0 @@
----
-name: summarize
-description: Summarize content from URLs, documents, or text. Use when the user wants a quick summary of something.
-argument-hint: [url or text]
-allowed-tools: web_fetch, read
----
-
-## Instructions
-
-Summarize the following content: $ARGUMENTS
-
-### Steps
-
-1. If the input is a URL, fetch the content using web_fetch
-2. If the input is a file path, read the file
-3. Otherwise, treat the input as text to summarize
-
-### Output Format
-
-Provide a summary with:
-- **TL;DR**: One sentence summary
-- **Key Points**: 3-5 bullet points of the main ideas
-- **Notable Details**: Any important specifics worth highlighting
-
-Keep the summary concise but comprehensive. Adapt the length based on the source content.

package/dist/.assistants/skills/skill-video/SKILL.md

@@ -1,113 +0,0 @@
----
-name: video
-description: Analyze video content — transcribe audio, extract key frames, answer questions about video. Supports local files and YouTube URLs.
-argument-hint: [video path or YouTube URL]
-allowed-tools: bash, read_audio, read
----
-
-## Instructions
-
-Analyze the video specified by: $ARGUMENTS
-
-### Step 1: Detect Input Type and Check Dependencies
-
-First, check if ffmpeg is available:
-
-```bash
-which ffmpeg || echo "MISSING: ffmpeg is required for video analysis. Install with: brew install ffmpeg"
-```
-
-Determine if the input is a YouTube URL or local file:
-- **YouTube URL** (contains youtube.com or youtu.be): Download with yt-dlp
-- **Local file**: Verify it exists
-
-For YouTube URLs, check yt-dlp is available:
-```bash
-which yt-dlp || echo "MISSING: yt-dlp is required for YouTube downloads. Install with: brew install yt-dlp"
-```
-
-### Step 2: Download (YouTube only)
-
-If the input is a YouTube URL, download the video:
-
-```bash
-mkdir -p /tmp/assistants-video
-yt-dlp -f "bestvideo[height<=720]+bestaudio/best[height<=720]" --merge-output-format mp4 -o "/tmp/assistants-video/%(id)s.%(ext)s" "URL_HERE"
-```
-
-Note the output filename for subsequent steps.
-
-### Step 3: Get Video Info
-
-```bash
-ffprobe -v quiet -print_format json -show_format -show_streams "/path/to/video" 2>/dev/null | head -100
-```
-
-This gives you duration, resolution, codec info.
-
-### Step 4: Extract and Transcribe Audio
-
-Extract audio from the video:
-```bash
-ffmpeg -i "/path/to/video" -vn -acodec pcm_s16le -ar 16000 -ac 1 "/tmp/assistants-video/audio.wav" -y 2>/dev/null
-```
-
-Then transcribe using the read_audio tool:
-- Use `read_audio` with path `/tmp/assistants-video/audio.wav`
-
-If the audio file is too large (>25MB), split it first:
-```bash
-ffmpeg -i "/tmp/assistants-video/audio.wav" -f segment -segment_time 300 -c copy "/tmp/assistants-video/audio_%03d.wav" -y 2>/dev/null
-```
-Then transcribe each segment separately.
-
-### Step 5: Extract Key Frames
-
-Extract frames at regular intervals (every 10 seconds, or every 300 frames):
-```bash
-ffmpeg -i "/path/to/video" -vf "fps=1/10" -frames:v 20 "/tmp/assistants-video/frame_%04d.jpg" -y 2>/dev/null
-```
-
-Adjust the fps and frame count based on video length:
-- Short video (<1 min): fps=1/5, max 12 frames
-- Medium video (1-10 min): fps=1/10, max 20 frames
-- Long video (>10 min): fps=1/30, max 20 frames
-
-### Step 6: Analyze Key Frames
-
-Use the `read` tool to view each extracted frame image. This sends the image to Claude's vision capability for analysis.
-
-For each frame, note:
-- What's visible in the frame
-- Any text, diagrams, or slides shown
-- Scene changes or transitions
-
-### Step 7: Synthesize Analysis
-
-Combine the transcription and visual analysis into a comprehensive report:
-
-**Output Format:**
-
-- **Video Info**: Duration, resolution, source
-- **Summary**: Overall summary of the video content
-- **Timeline**: Key moments with timestamps
-  - [0:00] - Description of opening
-  - [1:30] - Key point discussed
-  - etc.
-- **Transcript**: Full or summarized transcript
-- **Visual Elements**: Notable visual content (slides, diagrams, demonstrations)
-- **Key Takeaways**: Main points from the video
-
-### Step 8: Cleanup
-
-```bash
-rm -rf /tmp/assistants-video/
-```
-
-### Error Handling
-
-- If ffmpeg is not installed, inform the user: "ffmpeg is required for video analysis. Install with: brew install ffmpeg"
-- If yt-dlp is not installed (for YouTube): "yt-dlp is required for YouTube video downloads. Install with: brew install yt-dlp"
-- If the video file doesn't exist, inform the user with the correct path
-- If transcription fails (no ELEVENLABS_API_KEY), still proceed with visual analysis only
-- If the video is very long (>1 hour), suggest analyzing a specific time range

package/dist/.assistants/workflows/code-review.yml

@@ -1,23 +0,0 @@
-name: code-review
-description: Review staged or recent code changes for quality, bugs, and improvements
-version: "1.0"
-tags: [dev, review, quality]
-
-steps:
-  - name: Get diff
-    prompt: "Run 'git diff --staged' to see staged changes. If nothing is staged, run 'git diff' for unstaged changes. If still nothing, run 'git log -1 --format=%H' and 'git diff HEAD~1' to review the last commit."
-    allowed_tools: [bash]
-    output_variable: diff_output
-
-  - name: Identify changed files
-    prompt: "From the diff, list all changed files and categorize them: source code, tests, config, docs. Note the number of additions and deletions."
-    output_variable: file_summary
-
-  - name: Review for issues
-    prompt: "Review the code changes for: 1) Bugs or logic errors, 2) Security vulnerabilities (injection, auth issues), 3) Performance concerns, 4) Missing error handling, 5) Code style consistency. Be specific about line numbers and files."
-    allowed_tools: [read, grep]
-    output_variable: issues
-
-  - name: Suggest improvements
-    prompt: "Based on the review findings (${issues}), provide: 1) Critical issues that must be fixed, 2) Recommended improvements, 3) Optional nice-to-haves. For each suggestion, explain why and provide a code example if helpful."
-    output_variable: suggestions

package/dist/.assistants/workflows/content-creation.yml

@@ -1,22 +0,0 @@
-name: content-creation
-description: Create polished content through a structured outline, draft, and review process
-version: "1.0"
-tags: [writing, content]
-
-steps:
-  - name: Research and outline
-    prompt: "For the content topic '${topic}', create a detailed outline. Consider the target audience '${audience}' and format '${format}'. Research relevant information if needed. The outline should include main sections, key points, and supporting details."
-    allowed_tools: [web_search, web_fetch]
-    output_variable: outline
-
-  - name: Write first draft
-    prompt: "Using the outline (${outline}), write a complete first draft. Follow the specified format '${format}'. Write in a clear, engaging style appropriate for the audience '${audience}'. Include all sections from the outline with full content."
-    output_variable: draft
-
-  - name: Review and improve
-    prompt: "Review the draft (${draft}) for: 1) Clarity and flow, 2) Grammar and style, 3) Accuracy of information, 4) Engagement and readability, 5) Structure and organization. Provide specific revision suggestions."
-    output_variable: review
-
-  - name: Final polish
-    prompt: "Apply the review feedback (${review}) to create a final, polished version of the content. Ensure smooth transitions, strong opening and closing, consistent tone, and professional quality throughout."
-    output_variable: final_content

package/dist/.assistants/workflows/daily-standup.yml

@@ -1,24 +0,0 @@
-name: daily-standup
-description: Review recent git activity, summarize progress, and plan today's tasks
-version: "1.0"
-tags: [dev, daily, planning]
-
-steps:
-  - name: Review git history
-    prompt: "Run 'git log --oneline --since=yesterday --all' to see recent commits. Summarize what was worked on."
-    allowed_tools: [bash]
-    output_variable: git_summary
-
-  - name: Check current status
-    prompt: "Run 'git status' and 'git diff --stat' to see current changes in progress. What's the current state of the codebase?"
-    allowed_tools: [bash]
-    output_variable: current_status
-
-  - name: Review open tasks
-    prompt: "Check for any TODO comments, open issues, or pending tasks. Look at any task files or issue trackers available."
-    allowed_tools: [bash, read, glob, grep]
-    output_variable: open_tasks
-
-  - name: Generate standup report
-    prompt: "Based on the previous steps, generate a standup report with: 1) What was done yesterday (${git_summary}), 2) Current state (${current_status}), 3) Plan for today based on open items (${open_tasks}). Format as a clear, concise standup update."
-    output_variable: report

package/dist/.assistants/workflows/research.yml

@@ -1,23 +0,0 @@
-name: research
-description: Research a topic thoroughly by searching, gathering sources, and synthesizing findings
-version: "1.0"
-tags: [research, knowledge]
-
-steps:
-  - name: Initial search
-    prompt: "Search the web for '${topic}' using web_search. Find 5-10 relevant sources covering different aspects of the topic."
-    allowed_tools: [web_search]
-    output_variable: search_results
-
-  - name: Deep dive
-    prompt: "From the search results (${search_results}), fetch and read the top 3-5 most relevant sources using web_fetch. Extract key information, data points, and expert opinions."
-    allowed_tools: [web_fetch]
-    output_variable: source_content
-
-  - name: Cross-reference
-    prompt: "Compare information across the sources (${source_content}). Identify: 1) Points of agreement, 2) Conflicting information, 3) Unique insights from specific sources. Note which sources are most authoritative."
-    output_variable: analysis
-
-  - name: Synthesize report
-    prompt: "Create a comprehensive research report on '${topic}' based on the analysis (${analysis}). Include: Overview, Key Findings (with citations), Different Perspectives, Data/Statistics, Conclusions, and Source list with URLs. Be thorough but clear."
-    output_variable: report

package/dist/config/hooks.json

@@ -1,15 +0,0 @@
-{
-  "hooks": {
-    "SessionStart": [
-      {
-        "matcher": "*",
-        "hooks": [
-          {
-            "type": "command",
-            "command": "echo '{\"continue\": true, \"additionalContext\": \"Session started at '$(date)'\"}'"
-          }
-        ]
-      }
-    ]
-  }
-}