Generate targeted training challenges from edge cases and commit history.
npx skills add https://github.com/tyevans/tackline --skill challenge-gen
Please help me install this Agent Skill. Skill: challenge-gen Source: https://github.com/tyevans/tackline/tree/main/skills/workflows/challenge-gen Command: npx skills add https://github.com/tyevans/tackline --skill challenge-gen
This Skill automatically generates specific training challenges for AI agents, targeting their identified weaknesses to improve performance.
/diagnose-agent and identifying that an agent struggles with handling malformed user inputs, you can use challenge-gen to create specific scenarios that force the agent to practice robust input validation.Generate training challenges for the 'billing-agent'.
High-intent search queries and answers about installing and using this skill.
Targeted challenge generation creates edge-case and commit-replay scenarios by researching domain weaknesses and analyzing commit history to calibrate specific, grounded training data for AI agents.
Analyzing commit history enables commit-replay challenges, which recreate specific past scenarios and edge cases to force agents to practice resolving identified gaps and weaknesses in their performance.
After diagnosing agent weaknesses, challenge generation calibrates targeted edge-case scenarios to the identified struggle profile, forcing the agent to practice specific gaps like robust input validation.
Challenge generation supports both serial and parallel execution modes, allowing you to efficiently process multiple agent struggle profiles and generate targeted training scenarios at scale.
You need edge-case challenges when an agent struggles with specific scenarios like handling malformed user inputs, allowing you to create targeted training data that forces practice on robust input validation.
supertyrelleProfile agent strengths and weaknesses from learnings history and git signals.
MFS-codeMeasure page load performance and compare metrics against saved baselines.
supertyrelleGenerate targeted training challenges from agent struggle profiles with pipe-format metadata.
jsagirManually triggers domain decomposition analysis for a specified file via file-based material_id computation.