From f0efb25d91ea8a7165ed6aafd2b635d6163b6565 Mon Sep 17 00:00:00 2001 From: Magnus Hedemark Date: Sat, 11 Jul 2026 17:02:59 -0400 Subject: [PATCH] docs(skills): define completion and exit conditions --- AGENTS.md | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/AGENTS.md b/AGENTS.md index 09642f5..49995dc 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -127,7 +127,11 @@ Loading all 15 skills at session start (~6,000 lines, ~75KB) wastes context. Let When a skill body tells you to read a reference file only under specific conditions ("Read this if the API returns a 500"), do not read it proactively. Reference files are for specific edge cases, not general instruction. -### Validate Your Output +### Completion and Exit Conditions + +Skills that perform diagnosis, planning, or multi-step work must state when they are complete and when to stop. A valid exit condition is an observable artifact or a bounded escalation, such as: deliver the requested file, confirm the current setup is adequate, or stop after three non-converging diagnostic passes and report the evidence. + +## Validate Your Output When creating or modifying a skill in this repo, validate against the format: - `name` matches parent directory name