Help us improve
Share bugs, ideas, or general feedback.
From evolve-lite
Analyzes conversations to extract guidelines correcting agent reasoning: shortcuts for wasted steps, error preventions, user corrections for preferences.
npx claudepluginhub agenttoolkit/altk-evolve --plugin evolve-liteHow this skill is triggered — by the user, by Claude, or both
Slash command
/evolve-lite:learnThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
This skill analyzes the current conversation to extract guidelines that **correct the agent's reasoning chain**. A good guideline is one that, if known beforehand, would have led to a shorter or more correct execution. Only extract guidelines that fall into one of these three categories:
Captures agent mistakes, corrections, and discovered gotchas so they are not repeated. Use when: (1) a command or operation fails unexpectedly, (2) the user corrects the agent, (3) the agent discovers non-obvious behavior through debugging, (4) an API or tool behaves differently than expected, (5) a better approach is found for a recurring task. Also searches past learnings before starting tasks to avoid known pitfalls. Activate alongside the memory skill — they share sage-memory but serve different purposes (memory = codebase knowledge, self-learning = agent mistakes and gotchas). Also trigger on "sage review" or "review learnings" to curate and improve the learning database.
Captures high/medium/low confidence patterns from conversations to prevent repeating mistakes and preserve successes. Invoke proactively after corrections, praise, edge cases, or skill-heavy sessions.
Captures high/medium/low confidence learnings from conversations via triggers like corrections, praise, edge cases. Improves skills by preventing mistakes and preserving successes. Invoke proactively after 'no/wrong', 'perfect', or session ends.
Share bugs, ideas, or general feedback.
This skill analyzes the current conversation to extract guidelines that correct the agent's reasoning chain. A good guideline is one that, if known beforehand, would have led to a shorter or more correct execution. Only extract guidelines that fall into one of these three categories:
Do NOT extract guidelines that are:
DO extract guidelines for: environment-specific constraints discovered through errors (e.g., tools not installed, permissions blocked, packages unavailable) — these are not "known" until encountered in a specific environment.
Review the conversation and identify:
If none of these occurred, output zero entities. Not every conversation produces guidelines.
For each identified shortcut, error, or user correction, create one entity — up to 5 entities; output 0 when none qualify. If more candidates exist, keep only the highest-impact ones.
Principles:
State what to do, not what to avoid — frame as proactive recommendations
Triggers should be situational context, not failure conditions
For shortcuts, recommend the final working approach directly — eliminate trial-and-error by encoding the answer
For user corrections, use the user's own words — preserve the specific preference rather than generalizing it
Output entities as JSON and pipe to the save script. The type field must always be "guideline" — no other types are accepted.
echo '{
"entities": [
{
"content": "Proactive entity stating what TO DO",
"rationale": "Why this approach works better",
"type": "guideline",
"trigger": "Situational context when this applies"
}
]
}' | python3 ${CLAUDE_PLUGIN_ROOT}/skills/learn/scripts/save_entities.py
cat entities.json | python3 ${CLAUDE_PLUGIN_ROOT}/skills/learn/scripts/save_entities.py
python3 ${CLAUDE_PLUGIN_ROOT}/skills/learn/scripts/save_entities.py
# Then paste your JSON and press Ctrl+D
The script will:
.evolve/entities/){type}/ subdirectoriesBefore saving, review each entity against this checklist:
If any answer is no, drop the entity. Zero entities is a valid output.