Coding policy for Viktor Gamov's AI agents: language-agnostic quality rules, autonomous shipping discipline, and stack defaults for JVM, Swift, TypeScript, and Python
73
91%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Low
Low-risk findings worth noting
.beans.yml — work is tracked with the beans CLI, issues as Markdown files under .beans/beans-prime SessionStart hook loads the CLI's own usage guide (beans prime) at session start; that guide is the command reference and this rule is the discipline around it.beans/, no substitute tracker inventedbeans list --json --ready, beans list --json -S "<words>") or create one: beans create --json "<title>" -t <type> -d "<description>" -s in-progress## Summary of Changes section stating what was done## Reasons for Scrapping section, never silently deletedrules/boy-scout.md How to Apply: in a beans repo the "issue" is a bean)rules/commit-conventions.md One Change Per Commit still governs the shape)skills/release/SKILL.md Step 2)beans archive runs only when the user asks; a completed bean stays visible until thenETAG=$(beans show <id> --etag-only) then beans update <id> --if-match "$ETAG" ...; a conflict is resolved by re-reading, never by forcerules/agent-team-operation.md Task Ledger) is round evidence owned by the lead; a bean is the repo's record of the work — both exist in a beans repo, neither substitutes for the otherrules/architecture-decisions.md).beans.yml; a type is always given with -t.tessl-plugin
hooks
rules
skills
herdr-standup
herdr-teamlead
references
teamlead
templates
tests
onboard-repo
release
tests