Works with the AI tools you already use
Claude Code
Cursor
Codex CLI
GitHub Copilot
Gemini CLI+20 moreClaude CodeCursorCodex CLIGitHub CopilotGemini CLI
OpenClaw
Manus
Grok+17 agentsDiscussion Writer
Write a discussion that interprets results, owns limitations, and extracts real implications.
$7
discussion-writer
Example session with this skill installed
Use discussion-writer to write the discussion for my results, where my lightweight model (7M params) matched a 200M-param baseline on accuracy.
- Read your context and instructions
- Compiled the discussion-writer
Wrote a discussion that interprets the surprising small-model-matches-big-model result, addresses three threats to validity, and extracts implications for on-device deployment.
Discussion (draft)
Interpreting the key result
Our 7M-parameter model matching the 200M baseline on accuracy is surprising and worth scrutinizing. Two explanations are plausible: (1) the task is saturated at this accuracy ceiling, so capacity beyond roughly 7M adds no signal; or (2) the benchmark is too easy to discriminate models. The ablation (Section 4.3) supports (1): accuracy plateaus at 7M on all three benchmarks.
Limitations
First, we evaluate on English-only benchmarks, so cross-lingual behavior is unknown. Second, our models are trained once per seed (5 seeds), and variance is reported but not exhaustive. Third, efficiency is measured in FLOPs, not wall-clock latency on target hardware.
Threats to validity
Internal validity is strong (controlled training pipeline). External validity is limited by benchmark selection. Construct validity depends on whether accuracy captures user-perceived quality.
Implications
If capacity beyond 7M is wasted on these tasks, practitioners can deploy far cheaper models on-device without quality loss, which is a meaningful cost and privacy win.
Next steps
- Run
conclusion-writerto summarize these contributions and set up future work.
Connects securely to your tools. The creator never sees your data.
What you get
About this skill
The problem
Most discussions fail by repeating the results. A section that restates numbers without interpretation signals shallow analysis and invites "overclaiming" or "no critical insight" critiques. This skill writes one that actually interprets the findings.
What it does
- Explains surprising findings and why they happened.
- Acknowledges limitations honestly and addresses threats to validity.
- Extracts theoretical, practical, and methodological implications.
- Connects findings back to your research questions and the literature.
- Calibrates tone so you avoid overclaiming and undermining.
Why this beats prompting it yourself
General prompts rehash the results section or generate empty platitudes. This skill follows an interpretation-first structure, flags surprising results, considers alternative explanations, and preempts common reviewer critiques.
Use cases
- Writing the discussion section of a research paper.
- Interpreting results beyond surface-level description.
- Preparing a response to reviewer questions about interpretation.
- Extracting broader implications from findings.
- Writing a thesis discussion chapter.
Known limitations
The discussion is only as good as the results summary you provide; it cannot invent data or explain what was never measured. It interprets, but you must confirm that every claim stays within what the evidence supports.
Important
It interprets and contextualizes results; it does not run new analyses or produce new numbers. Run it after results are final, then hand the takeaway to conclusion-writer.
Why buy it
Use it for the discussion section or a response to reviewers. It is the second step of the Write Pack: interpret the results after you have positioned the paper.
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
30-day refund
Not a fit? Get your money back
Trust & safety
Security scanned
Verified clean 13 days ago
- Passed all security checks, Safe to install