New slides introduce LLM fundamentals before the Claude Code content for a PM
and beginner-technical audience:
- Slide 11 "One token at a time" — autoregressive generation (llm-basic.svg)
- Slide 12 "Tokens, not words" — tokenization basics (tokens.jpg)
- Slide 13 "Tokens in, tokens out" — combined view (llm-advanced.svg)
Renumbering: former slides 11–50 shifted to 14–53 by incrementing data-slide
attributes and updating SLIDE-N banner comments. data-level groupings
(agents/claude-md/skills/context/workflow) preserved. Total slide count
50 → 53. Asset path: ../assets/llm/.
Co-Authored-By: Claude <noreply@anthropic.com>
Animated SVG showing the same BPE-tokenized prompt from tokens.jpg (32 colored
subword tiles, e.g., "Anthropic" → "Anth"+"ropic", "Perplexity" →
"Per"+"plex"+"ity") feeding into the LLM and generating "Yes, they all use
BPE." token-by-token across 7 iterations. Combines tokenization and
autoregressive generation into one view.
Co-Authored-By: Claude <noreply@anthropic.com>
Screenshot of platform.openai.com/tokenizer showing the sentence "Does ChatGPT,
Claude, Anthropic, Llama, Mistral, Gemini, and Perplexity all use Byte-Pair
Encoding (BPE)?" tokenizing to 32 tokens / 105 characters. Visible tabs:
GPT-5.x & O1/3, GPT-4 & GPT-3.5 (legacy), GPT-3 (legacy) — illustrates that
different model generations use different tokenizers.
Co-Authored-By: Claude <noreply@anthropic.com>
Three-panel SVG (input context, LLM black box, predicted next token) with
7-iteration loop generating "The capital of Japan is Tokyo." from the prompt
"What is the capital of Japan?". Includes purple feedback loop showing each
predicted token appended back into the input.
Co-Authored-By: Claude <noreply@anthropic.com>
Insert the Matt Pocock entry between Lex Christopherson (GSD) and
Dani Avila (CC Templates) in the SUBSCRIBE -> Community section,
keeping the dev-workflow-creators cluster intact. Links to his X
handle and the Skills repo (51k stars, already referenced in the
videos table and the development-workflows table).
Co-Authored-By: Claude <noreply@anthropic.com>
Split the first bold tspan so only "Claude Pakistan" remains bold;
"community" now renders in normal weight while "WhatsApp" stays
bold. Aria-label and title are unchanged.
Co-Authored-By: Claude <noreply@anthropic.com>
- Promote the existing YouTube thumbnail under a new
`**🎬 Videos**` bold sub-heading
- Add a new `**📊 Presentations**` bold sub-heading with a
thumbnail (presentation-1.png, width 300 to match the video
thumbnail) linking to the GDG Kolachi "Claude Code & Gemini CLI"
deck on GitHub (tree view of
presentation/2026-04-25-gdg-kolachi-cli-claude-code-gemini/)
- Heading style follows the bold-inline `**Heading**` pattern
from the existing Specs & Documentation sub-section, with
emojis matching the repo's category-icon convention
Co-Authored-By: Claude <noreply@anthropic.com>
GDG Kolachi "Claude Code & Gemini CLI" presentation thumbnail used by
the new How to Use → Presentations sub-section in README.
Co-Authored-By: Claude <noreply@anthropic.com>
- Replace small how-to-use badge with new HD variant inside a `##`
heading and add explicit <a id="how-to-use"></a> anchor for
reliable linking from elsewhere in the README
- Expand the numbered list from 3 code-block lines to 7 markdown
bullets with parallel imperative leads and clickable cross-refs
to the Hooks repo, Agent Teams implementation, Tips section,
Subscribe section, and the new advanced-topics bullet pointing
at the Hot sub-table and the Ralph Wiggum implementation
- Add `[!TIP]` callout between the Boris Cherny block and the
CONCEPTS heading, directing readers to the How to Use section
- Rename three occurrences of shanraisshan/novel-llm-26 to
shanraisshan/ralph-wiggum-self-evolving-loop (CONCEPTS Hot row,
Development Workflows Others list, Tips and Tricks row)
- Refresh updated_with_Claude_Code badge timestamp to
May 02, 2026 3:17 PM PKT (version unchanged at v2.1.126)
Co-Authored-By: Claude <noreply@anthropic.com>
144x40 variant of how-to-use.svg with shimmer animation, matching the
orchestration-workflow-hd pattern but keeping the original green fill
(#2ea44f). Used by the How to Use section heading in README.
Co-Authored-By: Claude <noreply@anthropic.com>
Drops the !/root/second-most-starred.svg image from line 4. The next
commit deletes the SVG file itself.
Co-Authored-By: Claude <noreply@anthropic.com>
- Design vs Figma, Framer, Sketch, v0 (creative-tool tier disruption)
- Agent SDK vs LangChain, LangGraph, CrewAI, AutoGen, OpenAI Assistants
API (orchestration-framework tier disruption)
Inserted between Plan Mode and Skills/Plugins so the catch-all
"YC AI wrapper startups" row remains the table's editorial closer.
Co-Authored-By: Claude <noreply@anthropic.com>
- New row in Videos/Podcasts table for the 24 Apr 2026 workshop,
using the new community-matt.svg badge and linking to the local
transcript file
- New Planning/Specs tip on vertical slices vs horizontal phasing
(tracer bullets from Pragmatic Programmer); placed adjacent to
Dex's phase-wise plan tip as a contrasting refinement
- Bumped section count Planning/Specs (6) -> (7) and master count
TIPS AND TRICKS (82) -> (83)
Co-Authored-By: Claude <noreply@anthropic.com>
Full Walkthrough: Workflow for AI Coding (~2 hour conference workshop).
Covers grill-me alignment skill, smart/dumb zone theory, vertical
slices/tracer bullets, AFK Ralph loops, deep vs shallow modules
(Ousterhout), TDD red-green-refactor, and Sand Castle parallel agent
orchestration.
Organized into 17 topical sections instead of timestamped paragraphs
since the auto-transcript didn't include time anchors.
Co-Authored-By: Claude <noreply@anthropic.com>
Yellow community-palette badge (#f1c40f) for Matt Pocock, structured
identically to community-dex.svg/community-shayan.svg. Width 40px to
fit the 4-character "Matt" label.
Co-Authored-By: Claude <noreply@anthropic.com>
Pixel-art mascot raising a waving Pakistani flag, GitHub-flavored
50,000+ stars banner with shimmer/glow, tagline reset to vibe-coding-to-agentic-engineering line, tab title set to "50k stars on GitHub".
Co-Authored-By: Claude <noreply@anthropic.com>
Insert as row 2 alongside /workflows:development-workflows so the two top-level table-updater workflows sit together at the top, with the five best-practice/* drift-detector workflows stacked below as rows 3-7.
Co-Authored-By: Claude <noreply@anthropic.com>
- group the 4 affiliation banners (Claude for OSS, Community Ambassador, Certified Architect, Anthropic Academy) and the new WhatsApp community-invite banner under a new "## Extras" heading
- relocate Star History section from below Developed by to between the March 2026 trending image and Other Repos so the chart anchors a "trending → growth → other repos" narrative
Co-Authored-By: Claude <noreply@anthropic.com>
420x32 badge styled after badge-top-poster.svg pattern: WhatsApp green (#25D366) background, white WhatsApp logo on the left, Verdana text reading "Join Claude Pakistan community on WhatsApp" with the community name and platform name in bold.
Co-Authored-By: Claude <noreply@anthropic.com>
- keep only tags and named hyperlinks in the Description column across both tables
- drop /schedule link that pointed to /routines (text/URL mismatch surfaced once prose was removed); /schedule is still listed as a keyword in the Scheduled Tasks Location column and Routines has its own row
- move Routines next to Scheduled Tasks and Tasks so the umbrella → cloud surface → tracking trio reads in adjacency
Co-Authored-By: Claude <noreply@anthropic.com>
Appends a new dated entry (2026-05-01 03:29 PM PKT) for the Claude Code v2.1.126 settings update:
- 9 COMPLETE: version-bump, 3 new keys (preferredNotifChannel, ANTHROPIC_BEDROCK_SERVICE_TIER, CLAUDE_CODE_PROVIDER_MANAGED_BY_HOST), enriched bypassPermissions/defaultShell/spinnerTipsOverride descriptions, /config persistence note, Quick Reference example refresh
- 2 ON HOLD: OTEL_LOG_TOOL_DETAILS and OTEL_LOG_TOOL_CONTENT (changelog-only, recurring 16+ runs)
- 3 INVALID: spurious drift claims from research agent (autoSummaryEnabled non-existent, Agent(agent:name) prefix unverified, descriptive defaults flipped to "-" with no user benefit)
Co-Authored-By: Claude <noreply@anthropic.com>
Adds three new keys documented on the official settings/env-vars pages:
- preferredNotifChannel (Display Settings) — task-complete/permission notification channel with 7 enum values; default "auto" sends desktop notifications in iTerm2/Ghostty/Kitty
- ANTHROPIC_BEDROCK_SERVICE_TIER (Env Vars, v2.1.122) — Bedrock service tier (default/flex/priority) sent as X-Amzn-Bedrock-Service-Tier header
- CLAUDE_CODE_PROVIDER_MANAGED_BY_HOST (Env Vars, v2.1.126) — set by embedding host platforms; provider/auth env vars in settings.json are ignored when set
Enriches existing descriptions for v2.1.126 behavior changes:
- bypassPermissions exemption extended to .claude/, .git/, .vscode/, and shell config files
- defaultShell: PowerShell becomes primary shell when CLAUDE_CODE_USE_POWERSHELL_TOOL=1; PowerShell 7 detection covers Microsoft Store, MSI off-PATH, and .NET global tool installs
- spinnerTipsOverride.excludeDefault now also suppresses time-based tips (v2.1.121)
- /config now persists changes to ~/.claude/settings.json (v2.1.126)
Quick Reference example showcases the new keys.
Co-Authored-By: Claude <noreply@anthropic.com>
Shayan Rais
Claude Routine