add-didactic-comments
Skill13 starsChanged 51 days agoMITView on GitHub
Comment tagged files so the user learns the APIs they are studying, with a doc link in every comment.
In this repository5 other files it gives its agents.
- answer-as-typeSkill
- back-your-claimsSkill
- fill-the-gapsSkill
- just-do-itSkill
- no-bs-answerSkill
The file
We split this skill into the rules it gives an agent
9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
If that list is empty, go ahead with the skill.
- Distinctive: line 16
If it is not empty, stop.
- Very specific: line 32
IF topics is empty, THEN infer them from the imports of the tagged files, and say which ones you picked.
- Distinctive: line 33
IF a folder is tagged, THEN comment every source file under it.
- Distinctive: line 38
IF the anchor does not exist, THEN link the closest page that does and drop the anchor.
- Distinctive: line 51
Never edit code, names, or order.
- Very specific: line 52
Never delete a comment. Rewrite one only if it starts with AI and it is wrong.
- Distinctive: line 53
No temporal anchors, no change narration.
- Very specific: line 54
No comment restates the line.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Unstaged work
Files with unstaged changes, tracked and untracked:
!git status --porcelain --untracked-files=all | grep -E '^.[^ ]' || true
- If that list is empty, go ahead with the skill.
- If it is not empty, stop.
Reply with exactly this line and nothing else, no file list, no explanation, no preamble:
Stopping - unstaged files detected. Stage them and then invoke this skill again.Add didactic comments
The user tagged files and folders and named the topics they want to learn. Teach those topics inside those files. Change nothing but comments.
- Comment every call that belongs to
topics. Skip the rest of the code: plain python, obvious assignments, imports. - One comment per call, on the line above it.
- Say what the call does here, on this data. Not what the API does in general.
- End the comment with the doc URL of the exact symbol you commented, deep link included:
pandas.DataFrame.groupby, notpandas. - IF
topicsis empty, THEN infer them from the imports of the tagged files, and say which ones you picked. - IF a folder is tagged, THEN comment every source file under it.
Links
- Get the URL from the docs, never from memory. Fetch the page, confirm the anchor names the symbol, then paste the URL.
- IF the anchor does not exist, THEN link the closest page that does and drop the anchor. Never invent a fragment.
- Same symbol used twice in a file: link it once, at the first use.
Format
# {what this call does to this data}, {the argument that matters and why}
# https://{deep link to the symbol}
frame_by_year = sales.groupby("year")Rules
- Never edit code, names, or order. Comments only.
- Never delete a comment. Rewrite one only if it starts with
AIand it is wrong. - No temporal anchors, no change narration.
- No comment restates the line.
# read the fileabovepath.read_text()is noise; explain the encoding, the failure mode, the return type.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

