Back to search
Top Contributors
19
commits
A personal engineering handbook for the vibe coding era — collecting real mistakes from building AI agents
Stars
0
Forks
0
Watchers
0
Open Issues
0
Overall repository health assessment
No language data available
No package.json found
This might not be a Node.js project
19
commits
Add self-driving analogy to playbook Phase 2: why build small
2ea0047View on GitHubRefine two-layer agent mistake: context engineering before architecture
681991aView on GitHubAdd mistake: two-layer agent architecture where one agent + rules suffices
3cb126fView on GitHubAdd mistake: assuming one tool-calling protocol works across all agent CLIs
b4da00cView on GitHubAdd mistake: Telegram reply creates new thread, bot loses conversation context
dcd9a97View on GitHubAdd 3 mistakes: grounding vs persona contradiction, question feedback loop, validation-first pattern
7289c65View on GitHubAdd mistake: API default ordering puts low-quality duplicate first, agent trusts it
95d07eaView on GitHubAdd mistake: 3 post-processing bugs independently destroy URLs
2aed2c8View on GitHubAdd mistakes: ensure_ascii garbles CJK, text blobs lose URLs; fix earlier advice
fb914e7View on GitHubBoost: add exact field diff, pricing tier analysis, before/after output comparison
d03024aView on GitHubBoost: add concrete Google Places example to tool routing mistake
408b76dView on GitHubAdd mistake: incomplete tool audit logging makes failures unclassifiable
05d6d62View on GitHub