Abhinav Gupta·Sep 28Observability and Compliance in SkillsWhat actually stops an AI coding agent from running any skill it finds, and what tells you which ones it actually used?
Abhinav Gupta·Sep 27Deterministic vs Indeterministic — The Two Instruction Styles Every AI Agent NeedsWhy does the exact same agent setup nail a task on one run and go sideways on the next?
Abhinav Gupta·Sep 26What OpenTelemetry still can’t tell about the LLM callsWhen my code calls any LLM model, how do I actually know what happened inside the call?A response icon1
Abhinav Gupta·Sep 24Progressive Disclosure in Agentic AIWhy doesn’t an AI agent just load every tool it has access to, every time?A response icon1
Gavi Schneider·Aug 16📝 Send a Telegram Message in Pure PythonCredentials — Two env vars are required:
