v1.3.0

What's New

--update Mode Support

The skill now covers DocTest v1.9.0's --update flag for automatically fixing stale assertions:

• CLI Options: Added --update / -u flag documentation
• FIX Mode: New Step 3 — quick fix with --update for stale assertion values, auto-rewrites <!-- doctest: -->, <!-- doctest-json: -->, and // => with actual output
• Exit Codes: Updated to reflect --update mode behavior
• Classification Table: Added display output pattern

<!-- doctest-output --> Directive

New display-only output directive — blocks that get refreshed on --update but are never asserted against:

• Assertion Reference: Added Display Output Directive section with syntax and options (lines=N, tail=N)
• Classification Table: Added pattern for "output should be shown but not asserted"

Full Changelog: v1.2.0...v1.3.0

v1.2.0

What's New

MAKE-RUNNABLE Mode Rewrite

The MAKE-RUNNABLE mode has been completely rewritten with a new "runnable first, hide second" philosophy:

• 7-step workflow: Bootstrap Inventory → Classify → Make Runnable → Harden → Hide → Re-verify → Bootstrap Recommendations
• NEW vs EXISTING block workflows with detailed examples for each scenario
• Bootstrap profile support — leverage .doctest/ profiles instead of inline // [!code hide]
• file:N block targeting for fast iteration on individual blocks

Inline Assertion Strategy (// =>)

New cross-cutting strategy for proactively adding // => inline assertions to non-trivial expression lines:

• HARDEN step added to MAKE-RUNNABLE workflow (Step 4)
• Placement guide in assertions reference — clear rules for when to add vs skip
• Integrated into APPLY mode — Step 4 now considers // => for expression lines
• Principle: under-asserting is worse than over-asserting; users can remove unwanted assertions

Quality Examples

• APPLY mode: Converting a block with static output
• FIX mode: Fixing a dynamic output failure (hardcoded year → wildcard)
• MAKE-RUNNABLE: PriceCalculator API (NEW block), Laravel User model (EXISTING block), complex setup with hide:start/end
• HARDEN: Multi-step computation with intermediate value verification

Missing Feature Coverage

• // => dd() debug dump assertion type
• HTML comment attribute syntax (<!-- doctest-attr: ... -->)
• bootstrap="profile" attribute with profile discovery and execution order
• --parallel CLI option
• file:N block targeting syntax
• JSON reporter configuration
• Bootstrap recommendations for repeated patterns

Reference Updates

• assertions.md: Added Section 7 (Debug Dump) and Inline Assertion Placement Guide
• attributes.md: Added bootstrap attribute section and HTML Comment Attribute Syntax section
• make-runnable.md: Complete rewrite (133 → 305 lines) with philosophy, examples, and decision guide
• shiki.md: Synced to marketplace (was missing from previous release)

Full Changelog: v1.1.0...v1.2.0

v1.1.0 — MAKE-RUNNABLE Mode & Expanded Shiki Support

What's New

MAKE-RUNNABLE Mode

New mode that converts non-runnable documentation blocks into executable ones by adding hidden boilerplate with // [!code hide] markers. Hidden lines are invisible in rendered docs (VitePress/Shiki) but execute when DocTest runs the block.

Triggers: "make docs runnable", "add hidden setup"

Expanded Shiki Compatibility Reference

Full marker reference covering all 10+ Shiki markers with processing order documentation and per-marker behavior details.

Reference File Architecture

Detailed documentation split into focused reference files (reference/shiki.md, reference/make-runnable.md) keeping the main SKILL.md concise with summary + links.

Changes

• feat: Add MAKE-RUNNABLE as 5th skill mode (APPLY, VERIFY, FIX, REVIEW, MAKE-RUNNABLE)
• feat: Create reference/make-runnable.md — full MAKE-RUNNABLE workflow reference
• feat: Create reference/shiki.md — full Shiki marker reference (all 10+ markers)
• feat: Expand Shiki compatibility table in SKILL.md (3 → 10+ markers)
• docs: Update plugin.json, marketplace.json, README.md with new mode description