agentsclimarketplace

Lector

Skill fatihcatalkaya/lector-skill/skills/lector

Use when reviewing a near-final CS academic paper, master's thesis, or PhD thesis for writing quality and presentation before submission. Triggers on requests to "review", "proofread", "lector", or "check" a CS academic document.From its SKILL.md

Install
npx -y skills add fatihcatalkaya/lector-skill --skill lector

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.3 KB, as published. Nobody here has run it

Lector — Academic Writing Review

Near-final self-review of CS academic writing. Checks writing and presentation only — not scientific content, not academic integrity.

Language: American English
Discipline: Computer science only
Stage: Near-final (not early-draft restructuring)

Entry Flow

  1. Ask: Document type — paper / master's thesis / PhD thesis
  2. Ask: Document input — file path (you read it) or pasted text
  3. Read the full document
  4. Dispatch 4 subagents in parallel using the Agent tool:
    • Language agent → skills/lector/language-agent.md
    • Structure agent → skills/lector/structure-agent.md
    • Elements agent → skills/lector/elements-agent.md
    • References agent → skills/lector/references-agent.md
  5. Pass to each agent: full document text + document type
  6. Assign global IDs: concatenate the four subagent tables in fixed order — Language → Structure → Elements → References — and number rows sequentially starting at 1. IDs are unique across the whole report; do not reset per agent.
  7. Produce the consolidated report below.

Consolidated Report Format

Each agent gets its own titled section with a finding count and a single markdown table. If an agent reported no findings, keep the section header with (0 findings) and write No findings. instead of a table.

# Lector Review — [Document Title or Filename]
Document type: [Paper / Master's Thesis / PhD Thesis]
Reviewed: [date]

## Language & Style (N findings)
| ID | Location | Category | Finding | Zobel |
|----|----------|----------|---------|-------|
| 1  | §2.1 | Economy | "in order to" → "to" | Ch. 3 |
| 2  | §3   | Register | "we feel that the results are good" → "the results improve X by Y%" | Ch. 2 |

## Structure (N findings)
| ID | Location | Category | Finding | Zobel |
|----|----------|----------|---------|-------|
| 3  | Introduction | Ordering | motivation after method → move motivation before method | Ch. 7 |

## Elements (N findings)
| ID | Location | Category | Finding | Zobel |
|----|----------|----------|---------|-------|
| 4  | Fig. 2 | Caption | caption "Results" → state what is shown and the takeaway | Ch. 11 |

## References (N findings)
| ID | Location | Category | Finding | Zobel |
|----|----------|----------|---------|-------|
| 5  | [Smith24] | Missing entry | cited in §3 but absent from reference list → add bibliography entry | Ch. 6 |

## Summary
Total findings: N (Language: a | Structure: b | Elements: c | References: d)
Priority items: #2, #5, … — one-line reason per ID

Finding Format

Subagents emit rows of | Location | Category | Finding | Zobel |. The coordinator prepends an ID column with sequential global integers.

Out of Scope

  • Scientific content quality (methodology, evaluation soundness, contribution novelty)
  • Academic integrity / plagiarism detection
  • LaTeX/tool-specific formatting issues (overfull hboxes, etc.)
  • Non-CS academic disciplines
  • Early-draft structural rethinking

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.