Imlazy learn
Token-efficient tier-routed agent orchestrator 40+ skills, 19 agents, Obsidian vault. Works with Claude Code, OpenCode, Codex CLI.
npx -y skills add hnikoloski/imlazy --skill imlazy-learnAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 27 days oldThe repository was created 27 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 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.
What its author says it does
Copied from the file, not written here
Record router misclassifications to ~/.claude/imlazy/heuristics.log so the router adapts over time. Invoked by workflow skills on auto-escalation or on user feedback
SKILL.md
1.4 KB, 331 tokens by cl100k_base, as published. Nobody here has run it
learn
When to invoke
- Auto-escalation fired: the router classified Quick, work started, verification failed, escalated to Standard. Log it.
- User feedback: user signals classification was wrong ("this didn't need all that ceremony" or "this was bigger than you thought"). Log it.
What to write
Append one line to ~/.claude/imlazy/heuristics.log (create the directory if needed):
<ISO-8601-timestamp>\t<predicted-tier>\t<actual-tier>\t<task-summary-50chars>
Example:
2026-05-15T14:32:00Z quick standard add a /health endpoint
Implementation
mkdir -p ~/.claude/imlazy
printf '%s\t%s\t%s\t%.50s\n' \
"$(date -u +%Y-%m-%dT%H:%M:%SZ)" \
"$PREDICTED_TIER" \
"$ACTUAL_TIER" \
"$TASK_SUMMARY" \
>> ~/.claude/imlazy/heuristics.log
How the router uses it
lazy-router reads the last 20 lines of this file on every classification. Recent (≤30 days) misclassification patterns shift borderline classifications toward the tier that turned out correct in similar past cases.
What NOT to log
- Successful classifications — signal-to-noise only (log misclassifications only).
- Task content beyond 50 chars — truncation limits accidental leakage of sensitive task descriptions.
Gives 0 of the 12 instructions most learn study skills give in 331 tokens
Counted across 546 of the 573 authors here whose files we hold, read 2026-08-07
- calculate the zone of proximal development before teachingin 25 of 546, across 8 files
- produce self-contained HTML lessonsin 24 of 546, across 8 files
- record user preferences in a notes filein 23 of 546, across 5 files
- maintain a teaching workspace in the current directoryin 21 of 546, across 4 files
- find high-quality resources before writing lessonsin 19 of 546, across 5 files
- make lessons beautiful, short, and quickly completablein 19 of 546, across 3 files
- create reusable components for lessonsin 19 of 546, across 5 files
- create compressed reference documents for quick lookupin 19 of 546, across 3 files
- update the mission file and records upon mission changesin 16 of 546, across 2 files
- set min_dist to 0.0 for clustering preprocessingin 16 of 546, across 6 files
- populate the mission file before teachingin 15 of 546, across 1 file
- include interactive feedback loops in lessonsin 15 of 546, across 1 file
Said here and by no other author read
- log misclassifications on auto-escalation
- log misclassifications on user feedback
- create the target directory if needed
- append one line per misclassification
- use an iso-8601 timestamp
- record the predicted tier
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.