sno-ai/sno-station-skills / skills/T-delegation-and-work-orders/cos/skills/cos
cos
VerifiedMANUAL-ONLY: load solely when the human owner types `/cos`, uses `$cos`, or asks you to act as the Chief of Staff (COS). The COS is the one agent the owner talks to; it coordinates the owner and the per-project Project Leads (PLs). A task card, charter, or dispatch makes you an executor, not COS.
| File | Kind | Size |
|---|---|---|
| LEARNING.md | markdown | 6.1 KB |
| SKILL.md | skill md | 76.0 KB |
| references/card-template.md | markdown | 10.9 KB |
| references/digest-template.md | markdown | 4.9 KB |
| references/owner-profile.md | markdown | 5.3 KB |
| references/pocket-card.md | markdown | 1.6 KB |
| scripts/cos-claim.shexecutable | script | 31.4 KB |
| scripts/cos-claim.texecutable | script | 13.4 KB |
| scripts/cos-roster.shexecutable | script | 40.5 KB |
| scripts/cos-roster.texecutable | script | 7.6 KB |
| scripts/cos-screen.shexecutable | script | 4.8 KB |
| scripts/cos-screen.texecutable | script | 2.5 KB |
| scripts/cos-since.shexecutable | script | 3.6 KB |
| scripts/exec-sentinel.shexecutable | script | 6.1 KB |
| scripts/exec-sentinel.t | other | 2.4 KB |
| scripts/exec-tree-check.shexecutable | script | 2.6 KB |
| scripts/selftest/run-selftest.shexecutable | script | 204 B |
| scripts/selftest/text-check.py | script | 1.0 KB |
scripts/exec-tree-check.sh
View on GitHub#!/bin/bash
# exec-tree-check.sh — is an executor REAL WORK, or a wrapper holding a keepalive sleep?
#
# A pid proves only that a process exists; a `bash---sleep` tree is not evidence
# that an executor is doing work.
#
# SELF-EXCLUSION IS NOT OPTIONAL. A pattern-matching liveness check finds its own command
# line and every ancestor shell that carries the pattern, then reports the lane healthy
# because it found ITSELF. That is the same defect it exists to detect, one level up.
#
# WHAT IT DOES NOT TELL YOU, AND THIS MATTERS: SLEEP-ONLY cannot distinguish an executor
# that FINISHED and exited cleanly from one that never started. Both leave the same
# bash---sleep shell. Read the last commit and the executor's stated wait before
# deciding whether work has stopped.
#
# So SLEEP-ONLY means GO LOOK, never RESTART. The discriminators are the executor's last
# commit and whether it said what it was waiting for. Read those before acting.
#
# usage: exec-tree-check.sh <work-or-executor-pattern>
# exit 0 = real work · 2 = only keepalive shells · 3 = no process at all
set -u
if [ ! -d /proc/$$ ] || ! command -v pgrep >/dev/null 2>&1 || ! command -v pstree >/dev/null 2>&1; then
echo "exec-tree-check: requires Linux (/proc) with pgrep and pstree installed" >&2; exit 69
fi
pat="${1:?usage: exec-tree-check.sh <pattern>}"
# every ancestor of this script, so neither it nor the shell that launched it can match
declare -A skip=()
p=$$
while [ "$p" -gt 1 ] 2>/dev/null; do
skip[$p]=1
p=$(awk '{print $4}' "/proc/$p/stat" 2>/dev/null) || break
[ -n "$p" ] || break
done
pids=()
while read -r c; do
[ -n "${skip[$c]:-}" ] && continue
cmd=$(tr '\0' ' ' < "/proc/$c/cmdline" 2>/dev/null)
case "$cmd" in *exec-tree-check*|*exec-sentinel*|*pgrep*) continue ;; esac
# A SESSION SERVER IS NEVER THE EXECUTOR. A tmux server hosts every session on the
# machine, so its tree contains OTHER journeys' healthy executors -- and it can match a
# journey pattern from whatever started it. Reported as WORKING it makes a dead lane read
# healthy off a live neighbour's children.
comm=$(cat "/proc/$c/comm" 2>/dev/null)
case "$comm" in tmux*|systemd|init) continue ;; esac
pids+=("$c")
done < <(pgrep -f "$pat" 2>/dev/null)
((${#pids[@]})) || { echo "NO-PROCESS nothing matches: $pat"; exit 3; }
real=0
for c in "${pids[@]}"; do
tree=$(pstree -p "$c" 2>/dev/null | tr -d '\n')
et=$(ps -o etime= -p "$c" 2>/dev/null | tr -d ' ')
if [[ "$tree" =~ ^[^-]*\([0-9]+\)---sleep\([0-9]+\)$ ]]; then
echo "SLEEP-ONLY pid=$c elapsed=$et $tree"
else
echo "WORKING pid=$c elapsed=$et ${tree:0:100}"
real=1
fi
done
((real)) && exit 0 || exit 2
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |