Updates the read-only research agent so it detects (rather than
generates) prose in the CONCEPTS Description column. Description
Accuracy check now explicitly flags sentence-style summaries; Return
Format and critical rule #8 require recommended Description values to
contain badges + links only.
Co-Authored-By: Claude <noreply@anthropic.com>
Updates the executor command so future drift runs never reintroduce
prose into the CONCEPTS table Description column. Adds explicit
guidance in Phase 2 (Missing Concepts spec), Phase 3 (row template
now shows badge syntax), and as new critical rule #14.
Co-Authored-By: Claude <noreply@anthropic.com>
Agent self-update appending one Learnings bullet for the 2026-05-07 insert
of slide 9 "A horse. A model.": notes the stripped horse SVG composition,
the tightened viewBox (110 25 340 345 vs source's expanded -130 -50 780 450
that accommodated callouts), the etymology-line omission rationale, and a
regex-quirk for future integrity checks (CSS selector .slide[data-slide="1"]
matches naive data-slide="(\d+)" patterns — use <div[^>]+data-slide="(\d+)"
to scope to actual slide divs).
Co-Authored-By: Claude <noreply@anthropic.com>
Adds a Learnings entry to the GDG presentation agent capturing
(1) the etymology line just added to slide 13, (2) the actual
data-slide value of the horse-harness analogy slide in this deck so
future cross-deck syncs can reference it, and (3) a forward-looking
rule for cross-deck mirroring: when an edit lands on the sister
best-practices deck and the same slide exists here (i.e. inherited
from the original fork), evaluate whether to mirror — generic content
mirrors cleanly, event-specific content does not. Also notes that
slides should be located by content, not by trusting "slide N", since
post-fork drift can shift numbering between the two decks.
Co-Authored-By: Claude <noreply@anthropic.com>
Adds a Learnings entry for the 2026-04-30 slide-10 insert: documents
the styled-HTML approach chosen (no PNG asset existed for this dialog),
the slide-12 caption-strip pattern reuse, and a postmortem for the
sentinel-replacement collision where pre-resolving ##S10## to "11" before
running the n=11..48 loop double-incremented old-slide-10 to "12". Adds
two forward-looking rules: (1) use distinct sentinel prefixes or
single-pass resolution when mixing pre-resolved and loop-resolved slides,
and (2) check raw file content for literal apostrophes vs HTML entities
when pattern-matching comment strings.
Co-Authored-By: Claude <noreply@anthropic.com>
Appends a Learnings entry to the presentation-claude-gemini agent capturing
the 2026-04-30 removal of "How the Harness Reaches the World" from the GDG
deck and the deliberate-divergence framing — the slide lives on in the
forked claude-code-best-practice deck but is gone from the GDG historical
record because it was added post-event.
Future runs by this agent will know why the two decks differ on this slide
and won't try to "reconcile" them.
Co-Authored-By: Claude <noreply@anthropic.com>
Updates the per-presentation delegation rule to route requests on the new
"Claude Code Best Practice" deck to the new presentation-claude-code agent.
Previously the table had only two decks (vibe-coding, claude-gemini); now
all three are listed with explicit path → agent mapping and an invocation
example.
Adds clarification that "the main presentation" or "the best-practice deck"
generally refers to the new claude-code agent.
Co-Authored-By: Claude <noreply@anthropic.com>
New per-deck agent owns presentation/claude-code-best-practice/index.html.
Mirrors the per-presentation pattern of presentation-claude-gemini and
presentation-vibe-coding — each deck has its own focused knowledge base
that self-evolves without cross-contaminating the others.
Seeded with structural facts about this deck specifically (no journey bar,
no hardcoded goToSlide(N) calls, simpler than the GDG fork) and a fresh
Learnings section starting at the 2026-04-30 fork date — GDG agent's 25+
historical entries deliberately not copied since most are tied to features
this deck doesn't have.
Co-Authored-By: Claude <noreply@anthropic.com>
Adds the fff3b0/ddf4ff color split to the Workflow column spec, asks
research agents to mark each step as (top) or (sub) in the structured
report, updates the Phase 3 execute step to map those markers to badge
colors, and makes the post-table legend mandatory via new Rule 11.
Co-Authored-By: Claude <noreply@anthropic.com>
Updates the /workflows:development-workflows command to require
HTML <img align="middle"> badges (so arrows align with badge
centers) and to forbid parenthetical sub-loops or English
qualifiers in the Workflow chain. Inner loop steps must be
listed as flat siblings.
Co-Authored-By: Claude <noreply@anthropic.com>
Removes the Plan-implementation research item from both agent prompts,
strips the PLAN line from the structured report format, and prunes the
Plan-link rules from Phase 0/2/3 and the Rules section. Column order
in the spec now matches the README: Name | ★ | Workflow | a | c | s.
Co-Authored-By: Claude <noreply@anthropic.com>
Two agent files both declared `name: time-agent` with different implementations:
- `.claude/agents/time-agent.md` (root): PKT, UTC+5 (Pakistan Standard Time)
- `agent-teams/.claude/agents/time-agent.md` (agent-teams): Dubai GST, UTC+4
When both scopes are active, Claude Code may invoke the wrong time-agent —
serving PKT time when a Dubai command expects GST, or vice versa.
Renamed the root agent's `name` from `time-agent` to `time-agent-pkt` so
the two agents are unambiguous. The agent-teams variant retains `name: time-agent`
since `agent-teams/.claude/commands/time-orchestrator.md` explicitly invokes it
by that name for Dubai time display.
Co-Authored-By: Claude Code <noreply@anthropic.com>
The '# Glob: presentation/**' markdown heading was not parsed as scoping
metadata — Claude Code only reads YAML frontmatter with paths: for
path-specific rule loading. Without proper frontmatter, the file loaded
into every session instead of only presentation/** work. Now correctly
scoped via paths: ["presentation/**"] with H1 'Presentation Delegation'.
Co-Authored-By: Claude <noreply@anthropic.com>
The '# Glob: **/*.md' markdown heading was a cargo-cult label — Claude Code
only parses YAML frontmatter with paths: for path-scoping. Without it, the
file loaded unconditionally on every session. Now correctly scoped via
paths: ["**/*.md"] and H1 retitled to 'Markdown Docs' for human readers.
Co-Authored-By: Claude <noreply@anthropic.com>
Documents the below-fold legend technique: min-height expansion, scroll affordance,
window.scrollTo(0,0) in showSlide(), and why no CSS ancestor changes are needed.
Co-Authored-By: Claude <noreply@anthropic.com>
Documents the color-coded definition table technique: shrink chip-cloud
container to free vertical space, add 2-col grid below with matching hex
term colors, 0.82rem rows, 9 items per column balancing color tiers.
Co-Authored-By: Claude <noreply@anthropic.com>
Self-evolve log entries from the Apr 18 "What is Claude Code?" slide
insertion: descending-sed renumber method, three-column inline-grid layout
choice, post-insertion section-divider positions (Agents=8 ... Workflow=34),
and the sequencing rule that Edit must run after sed completes — parallel
execution causes "file modified since read" errors.
Co-Authored-By: Claude <noreply@anthropic.com>
Move disableAllHooks next to the opening brace so the master toggle
is visible at the top instead of buried below env and
enableAllProjectMcpServers. Behavior unchanged (remains false).
Co-Authored-By: Claude <noreply@anthropic.com>
Add Write, Edit, Agent, NotebookEdit, and mcp__* to the allowlist so
the research agent can author reports and delegate to subagents
without per-tool permission prompts.
Co-Authored-By: Claude <noreply@anthropic.com>
Added mandatory learning: never run destructive git ops (checkout --, restore,
reset --hard) on uncommitted work. Added commit-per-milestone protocol, Write-tool
vs sed preference for large restructures, and full rebuild summary for the
2026-04-18 weather-reporter 37-slide deck.
Co-Authored-By: Claude <noreply@anthropic.com>
Shayan Rais
claude[bot]