feat: update SKILL.md metadata format and eval reports across all skills

Rename author field to skill-author, remove redundant source links,
add content improvements to multiple skills, and refresh eval report results.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: GiaoLee

awesome-med-research-skills/Academic Writing/author-response-builder/SKILL.md

@@ -2,9 +2,8 @@
 name: author-response-builder
 description: Turns reviewer comments into structured, professional point-by-point responses linked to manuscript revisions, clarifications, rebuttals, and additional analyses.
 license: MIT
-author: aipoch
+skill-author: AIPOCH
 ---
-> **Source**: [https://github.com/aipoch/medical-research-skills](https://github.com/aipoch/medical-research-skills)
 
 # Author Response Builder
 
@@ -141,6 +140,10 @@ This skill should:
 If the user provides only fragments of reviewer comments, vague revision notes, or no clear information about what has actually changed in the manuscript, do not immediately produce a full point-by-point response.
 First explain what is missing, ask focused follow-up questions, or recommend uploading the full review package and revision notes.
 
+**Constructive pivot for incomplete revisions:** If the user has not completed revisions, do not simply refuse. Instead offer: "I can draft provisional responses for any revisions you can describe now, noting that final revision-linkage text should be confirmed once revisions are complete. Which comments have you already addressed?" This keeps the interaction productive without fabricating completed changes.
+
+**Editor letter format:** Editor decision letters should be addressed as a single block response unless the editor letter contains enumerated action items. Tone should be slightly more formal than reviewer responses, and the opening should directly acknowledge the editorial decision.
+
 ### Step 2 — Identify the response unit
 Determine whether the response should be built:
 - comment by comment,
@@ -165,6 +168,13 @@ Check whether the response should point to:
 - limitation statement,
 - or no manuscript change with a transparent explanation.
 
+### Step 4.5 — Tiered Output Mode
+
+Apply the following output mode based on input complexity:
+
+- **Simple mode** (2 or fewer comments, all fully resolved acceptances): Combine Sections F, G, and H into a single brief "Notes" block. Reserve the full 8-section structure for mixed-mode, partially-resolved, or rebuttal-containing responses.
+- **Complex mode** (5 or more comments): Section C must include a mode-distribution count (e.g., "3 acceptances, 2 explanations, 1 rebuttal, 1 additional-analysis, 1 partial-with-limit") rather than a generic "mixed" label.
+
 ### Step 5 — Draft the point-by-point response
 Construct each reply so that it:
 - acknowledges the comment,
@@ -233,6 +243,10 @@ When helpful, recommend uploading reviewer comments, editor letter, revision str
 - Do not overuse empty courtesy formulas.
 - Do not produce a confident full response package when the actual revision status is still too incomplete.
 
+## Editorial Consequence Rule
+
+When refusing to produce dismissive or reviewer-targeting language, explain the downstream editorial consequence: editors routinely interpret dismissive responses as author inflexibility and often rule in favor of the reviewer, increasing rejection probability. Frame the alternative bounded rebuttal as the strategically stronger choice — not just the polite one.
+
 ## Hard Rules
 
 1. **Do not invent completed manuscript changes, analyses, experiments, or figure revisions.**

awesome-med-research-skills/Academic Writing/author-response-builder/eval_report_author-response-builder_result.json

@@ -2,9 +2,8 @@
 name: claim-strength-calibrator
 description: Calibrates manuscript claim strength so wording matches the actual evidence level, study design, and validation status.
 license: MIT
-author: aipoch
+skill-author: AIPOCH
 ---
-> **Source**: [https://github.com/aipoch/medical-research-skills](https://github.com/aipoch/medical-research-skills)
 
 # Claim Strength Calibrator
 

awesome-med-research-skills/Academic Writing/claim-strength-calibrator/SKILL.md

@@ -2,9 +2,8 @@
 name: conference-abstract-writer
 description: Condenses a full study into conference-submission abstract format. Use when adapting a manuscript abstract or study summary to meet a specific conference's word limit, structured format (Background/Methods/Results/Conclusion), character limits, or required section headings. Also triggers on "adapt my abstract for [conference]", "shorten my abstract to 250 words", "reformat for ASCO/ASGCT/SfN/AACR", "I need a conference abstract", or "cut my abstract to fit the word limit".
 license: MIT
-author: aipoch
+skill-author: AIPOCH
 ---
-> **Source**: [https://github.com/aipoch/medical-research-skills](https://github.com/aipoch/medical-research-skills)
 
 # Conference Abstract Adaptor
 
@@ -32,13 +31,15 @@ Out-of-scope:
 
 ## Supported Conference Formats
 
+> ⚠️ **Conference requirements change annually.** Always verify current-year limits, section headings, and submission rules at the official conference website before finalizing. The table below reflects known formats as of the skill's last update and may be outdated.
+
 | Conference | Limit | Format |
 |---|---|---|
 | **ASGCT** | 250 words | Structured: Background / Methods / Results / Conclusion |
 | **ASCO** | 260 words | Structured: Background / Methods / Results / Conclusions |
 | **AACR** | 300 words | Structured: Background / Methods / Results / Conclusions |
 | **ASM** | 300 words | Single-paragraph or structured (conference-dependent) |
-| **SfN** | 2,000 characters | Single paragraph (no headings) |
+| **SfN** | 2,000 characters (including spaces) | Single paragraph (no headings) |
 | **ESC / AHA / ACC** | 250–350 words | Structured (verify current year requirements) |
 | **Custom** | User-specified | User-specified |
 
@@ -82,7 +83,9 @@ When cutting to fit the word limit:
 After adapting, provide:
 1. The adapted abstract, fully formatted
 2. The word count (or character count) with a clear note: `Word count: X / [limit]`
+   - **For character-limited formats (e.g., SfN):** state the count as `[N] characters including spaces`. SfN and most character-limited conferences count spaces. Users should verify using a tool that counts spaces (e.g., Python `len()` or Word character count with spaces enabled).
 3. A one-line note on any content that was cut to fit the limit, so the user can decide whether to restore it
+4. **Word-limit conflict resolution:** If preserving the primary result, sample size N, and conclusion still exceeds the word limit after aggressive cutting, notify the user: "Preserving required elements (primary result, N, conclusion) results in [X] words vs. [limit]. Please indicate which secondary elements to deprioritize." Do not silently cut required elements.
 
 ### Step 5 — Final Check