friction
Skill2 starsChanged 5 months agoMITView on GitHub
Capture agentic friction in real time as it happens during a session — corrections, surprises, tool gaps, abandoned approaches. Use mid-session when you hit friction or get redirected. Produces a structured friction log in markdown. Invoke with /skill:friction.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- manusco/resonancetheir own version, first seen Oct 22, 2026
- gabo963/claude-skillstheir own version, first seen Oct 15, 2026
- ananmouaz/frictiontheir own version, first seen Oct 13, 2026
- udaysharmadev/Skillstheir own version, first seen Sep 17, 2026
- jorekai/skillstheir own version, first seen Sep 7, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- manusco/resonance★ 38same name
- udaysharmadev/Skills★ 12same name
- ananmouaz/friction★ 0same name
- gabo963/claude-skills★ 0same name
- jorekai/skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository2 other files it gives its agents.
- retroSkill
- self-reviewSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Don't wait for end-of-session.
- Distinctive: line 51
Keep rows terse: If a friction event needs more than one line, it's probably worth a retro item, not a friction row.
- Distinctive: line 57
If retro isn't run this session, the log still has value.
“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Friction — Real-Time Friction Capture
Purpose: Friction is fuzzy by end-of-session. Capture it while it's hot, with the context that caused it. This log is the primary input to a post-session retrospective.
Friction here means: anything that made the session harder than it needed to be — a wrong assumption, a tool that didn't work as expected, an unclear instruction, a retry, an abandoned approach, a place you hedged or got corrected.
When to use
- The user corrects or redirects you ("no, I meant...", "actually do it this way")
- You retry something that should have worked the first time
- You abandon an approach mid-stream
- A tool/skill behaves unexpectedly
- You notice yourself hedging or uncertain
- The user says "friction:" as a prefix to log something
Don't wait for end-of-session. Log as it happens — one line per entry is enough. The point is to capture the moment, not write prose.
Workflow
Capture
Append to the friction log. Use a single file per session.
File: .agent-friction/friction-<YYYY-MM-DD>.md (configure the directory to suit your setup; ensure it's gitignored if these are working notes, or committed if you want them shared)
Create if missing:
# Friction Log — <date>
| # | Time | Severity | What | Context | Apparent cause |
|---|------|----------|------|---------|----------------|Then append a row per friction event:
| Column | Guidance |
| Time | HH:MM (approx) |
| Severity | 🟢 minor (slowed me) / 🟡 notable (wrong direction) / 🔴 material (had to redo or get help) |
| What | One phrase — what happened |
| Context | One phrase — what were you trying to do |
| Apparent cause | One phrase — your first guess at why (retro goes deeper) |
Keep rows terse. If a friction event needs more than one line, it's probably worth a retro item, not a friction row. Capture the seed; let retro grow it.
Close the log at end of session
The retrospective skill reads this file. After retro runs, the friction log is consumed — leave it as historical record (don't delete).
If retro isn't run this session, the log still has value; it stays.
What this is NOT
- Not a TODO list. Friction rows aren't action items. Retro turns them into action items. The log is raw observation.
- Not a venting log. "I'm frustrated" isn't friction. "I'm frustrated because I assumed X and had to redo Y" is.
Relationship to other skills
- Feeds the retrospective skill — retro's "where did I hedge/fumble/get corrected" question draws directly from this log.
- Feeds self-review — if friction was about your own work quality, self-review may consume it.
v0.1.0: Extracted from a collaboration hub and generalized. Inspired by aurorascharff/agent-friction-skill (passive end-of-session variant) and the friction-detection phase of giannimassi/agent-retro. Severity coding 🟢🟡🔴 from agent-friction-skill.
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.