Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/timoncool/dream-skill/wakenpx skills add timoncool/dream-skill --skill wakegit clone --depth 1 https://github.com/timoncool/dream-skillWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/timoncool/dream-skill/wake)<a href="https://agentmods.dev/skills/timoncool/dream-skill/wake"><img src="https://agentmods.dev/badge/skills/timoncool/dream-skill/wake.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00265 | $0.06422 |
| Opus 5 | $0.00133 | $0.03211 |
| Sonnet 5 | $0.00053 | $0.01284 |
| Haiku 4.5 | $0.00026 | $0.00642 |
Grade C, and why
wake scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf "$LOCK_DIR" && mkdir "$LOCK_DIR" && echo $$ > "$LOCK_DIR/pid" How it starts
The opening of the file, as written. The whole thing — 420 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Wake — применение выбранного из dream
После того как dream сделал отчёт и пользователь отметил галочками — wake подхватывает выбор и применяет.
Принцип безопасности
wake НИКОГДА не изобретает что применять. Источники выбора (по приоритету):
- JSON файл
DREAM-CHOICES-<date>.json(из HTML кнопкой Save, либо создан валидатором в auto mode — помечен"auto": true, см. dream SKILL.md «Auto mode») - Аргументы skill'а —
wake M1,M3,N2илиwake all
Если ни один не задан — спрашивает пользователя что выбрать, не делает ничего.
Workflow
Phase 1 — Lock + найти отчёт и выбор
Сначала lock против race (как в dream/SKILL.md Phase 0). Если две сессии одновременно запустят wake на одном отчёте — Edit'ы перезатрут друг друга, MEMORY.md превратится в кашу.
LOCK_DIR="$CWD_BASH/.wake-lock"
if mkdir "$LOCK_DIR" 2>/dev/null; then
echo $$ > "$LOCK_DIR/pid"
else
LOCK_AGE=$(( $(date +%s) - $(stat -c %Y "$LOCK_DIR" 2>/dev/null || stat -f %m "$LOCK_DIR") ))
if [ "$LOCK_AGE" -gt 3600 ]; then
rm -rf "$LOCK_DIR" && mkdir "$LOCK_DIR" && echo $$ > "$LOCK_DIR/pid"
else
echo "WAKE ALREADY RUNNING (lock age ${LOCK_AGE}s) — abort"
exit 0
fi
fi
# В финале Phase 5: rm -rf "$LOCK_DIR"
Сначала вычислить пути (как в dream/SKILL.md — _BASH для bash команд, _WIN для Python):
# Compute paths (Win11 Git Bash)
CWD_BASH=$(pwd) # /d/Projects/TEMP
CWD_WIN=$(pwd -W 2>/dev/null || pwd) # D:/Projects/TEMP (на не-Windows fallback)
SLUG=$(echo "$CWD_WIN" | sed 's|[:/]|-|g') # D--Projects-TEMP
MEMORY_DIR_BASH="$HOME/.claude/projects/$SLUG/memory"
MEMORY_DIR_WIN=$(cygpath -w "$MEMORY_DIR_BASH" 2>/dev/null | sed 's|\\|/|g' || echo "$MEMORY_DIR_BASH")
test -d "$MEMORY_DIR_BASH" && echo "OK $MEMORY_DIR_WIN" || { echo "MEMORY DIR NOT FOUND — wake cannot proceed without it"; }
# 1. Find latest dream report MD in cwd
LATEST_REPORT=$(ls "$CWD_BASH"/DREAM-REPORT-*.md 2>/dev/null | sort -r | head -1)
if [ -z "$LATEST_REPORT" ]; then
echo "NO REPORT FOUND — run dream skill first to generate DREAM-REPORT-<date>.md"
# Don't exit — Claude reads this output and stops the workflow itself
else
echo "Report: $LATEST_REPORT"
fi
# Extract date from filename: DREAM-REPORT-2026-05-02.md → 2026-05-02
REPORT_DATE=$(basename "$LATEST_REPORT" | sed 's|DREAM-REPORT-||; s|\.md||')
# 2. Find choices JSON (priority: cwd → ~/Downloads/ → ~/Desktop/)
CHOICES=""
for path in "$CWD_BASH/DREAM-CHOICES-$REPORT_DATE.json" "$HOME/Downloads/DREAM-CHOICES-$REPORT_DATE.json" "$HOME/Desktop/DREAM-CHOICES-$REPORT_DATE.json"; do
if [ -f "$path" ]; then CHOICES="$path"; break; fi
done
# Fallback: latest CHOICES regardless of date
if [ -z "$CHOICES" ]; then
CHOICES=$(ls -t "$HOME/Downloads/DREAM-CHOICES-"*.json "$CWD_BASH/DREAM-CHOICES-"*.json 2>/dev/null | head -1)
fi
echo "Choices: ${CHOICES:-NOT FOUND}"
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 420 lines · 0 tokens per session scan C 911d17b87271
wake is a skill published in the GitHub repository timoncool/dream-skill (3 stars, last pushed 10d ago), licensed MIT. It adds 265 tokens to every session and 6,422 once invoked, about $0.0013 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
metabot
Unified MetaBot CLI for personal Memory, Skill Hub, durable Agent Bus messaging, agent registry, Agent Teams, T5T, scheduling, and bridge runtime operations.
oak
Find out what happened, what was decided, and what depends on what in your codebase. Use this skill whenever you need to: recall past decisions or discussions ("what did we decide about X?"), check what might break before refactoring ("what depends on this module?"), find conceptually similar code that grep would miss…
takenotes
Saves durable knowledge into permanent storage — typed memory, CLAUDE.md, or a docs/ file — and corrects anything already stored that has gone stale. Handles both a single fact and a whole-session harvest. Use when the user says "takenotes", asks you to remember or record something, asks you to save what was learned…
putdown
Writes a session handoff file that a fresh Claude Code session reads via /pickup, harvests durable knowledge into memory and CLAUDE.md (via /takenotes when installed), then commits and pushes all session work. Use when the user says "putdown", when ending or stepping away from a working session, or when the context…
pickup
Loads the most recent putdown handoff file for the current project and primes the session with full prior context. Use when the user says "pickup", asks "where were we" after opening a fresh window, or wants to resume work from a prior session that ended with /putdown. Do NOT use for reopening a previous Claude…
AgentDB Memory Patterns
Implement persistent memory patterns for AI agents using AgentDB. Includes session memory, long-term storage, pattern learning, and context management. Use when building stateful agents, chat systems, or intelligent assistants.