Du Dum
Separate expensive observation from cheap decision-making in autonomous agent loops using a two-clock architecture. A fast clock accumulates data into a digest file; a slow clock reads the digest and acts only when something is pending. Idle cycles cost nothing because the action clock returns immediately after reading an empty digest. Use when building autonomous agents that must observe continuously but can only afford to act occasionally, when API or LLM costs dominate and most cycles have nothing to do, when designing cron-based agent architectures with observation and action phases, or wh
Install
npx skills add pjt222/agent-almanac --skill du-dumgit clone --depth 1 https://github.com/pjt222/agent-almanac.git /tmp/agent-almanac
mkdir -p ~/.claude/skills
cp -r /tmp/agent-almanac/i18n/caveman-lite/skills/du-dum ~/.claude/skills/du-dumPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Du-Dum: Batch-Then-Act Pattern Separate observation from action using two clocks running at different frequencies. The fast clock (analysis) collects data cheaply and writes a compact digest. The slow clock (action) reads the digest and decides whether to act. If the digest says nothing pending, the action clock exits immediately -- zero cost for idle cycles. The name comes from the heartbeat rhythm: du-dum, du-dum. The first beat (du) observes; the second beat (dum) acts. Most of the time, only the first beat fires. ## When to Use - Building autonomous agents that run on a budget and must observe more often than they act - An existing heartbeat loop calls the LLM every tick, even when nothing has changed - Observation is cheap (API reads, file parsing, log scanning) but action is expensive (LLM calls, write operations, notifications) - You need decoupled failure: if observation fails, the last good digest should still be valid for the action clock - Designing cron-based agent architectures where analysis and action run as separate jobs ## Inputs - **Required**: List of data sources the fast clock should observe (APIs, files, logs, feeds) - **Required**: Action the slow clock should take when the digest indicates pending work - **Optional**: Fast clock interval (default: every 4 hours) - **Optional**: Slow clock interval (default: once per day) - **Optional**: Cost ceiling per day (to validate the clock configuration)
Details
- Skill name
du-dum- Repository
- pjt222/agent-almanac
- Path
i18n/caveman-lite/skills/du-dum/SKILL.md- License
- MIT
- Repo updated
- 18 November 2026
- Listed in
- BehiSecc/awesome-claude-skills
FAQ
What does the Du Dum skill do?
Separate expensive observation from cheap decision-making in autonomous agent loops using a two-clock architecture. A fast clock accumulates data into a digest file; a slow clock reads the digest and acts only when something is pending. Idle cycles cost nothing because the action clock returns immediately after reading an empty digest. Use when building autonomous agents that must observe continuously but can only afford to act occasionally, when API or LLM costs dominate and most cycles have nothing to do, when designing cron-based agent architectures with observation and action phases, or wh
How do I install the Du Dum skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add pjt222/agent-almanac --skill du-dum. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Du Dum popular?
It lives in pjt222/agent-almanac, which has 45 GitHub stars and 4 forks, and is included in 1 curated skill list.
Related skills
Adapt Architecture
45pjt222/agent-almanac
Execute structural metamorphosis using strangler fig migration, chrysalis phases, and interface preservation. Covers transformation planning, parallel running, progressive cutover, rollback design, and post-metamorphosis stabilization for system architecture evolution. Use when assess-form has classified the system as READY for transformation, when migrating from monolith to microservices, when replacing a core subsystem while dependents continue operating, or when any architectural change must be gradual rather than big-bang.
Adaptic
45pjt222/agent-almanac
Master skill composing the 5-step synoptic cycle for panoramic synthesis across multiple domains. Orchestrates meditate, expand-awareness, observe, awareness, integrate-gestalt, and express-insight into a coherent process that produces unified understanding rather than sequential compromise. Use when a problem genuinely spans 3+ domains and the interactions between domains matter more than depth in any one, when sequential analysis feels like compromise rather than integration, or before major architectural decisions affecting multiple stakeholders.
Add Puzzle Type
45pjt222/agent-almanac
Scaffold a new puzzle type across all 10+ pipeline integration points in jigsawR. Creates the core puzzle module, wires it into the unified pipeline (generation, positioning, rendering, adjacency), adds ggpuzzle geom/stat layers, updates DESCRIPTION and config.yml, extends the Shiny app, and creates a comprehensive test suite. Use when adding a completely new puzzle type to the package or following the 10-point integration checklist to ensure nothing is missed end-to-end.
Evolve Agent
45pjt222/agent-almanac
Evolve an existing agent definition by refining its persona in-place or creating an advanced variant. Covers assessing the current agent against best practices, gathering evolution requirements, choosing scope (refinement vs. variant), applying changes to skills, tools, capabilities, and limitations, updating version metadata, and synchronizing the registry and cross-references. Use when an agent's skills list is outdated, user feedback reveals capability gaps, tool requirements have changed, an advanced variant is needed alongside the original, or the agent's scope needs sharpening after real
Dream
45pjt222/agent-almanac
AI unconstrained creative exploration — drop structure to let associations form freely. The one skill that inverts standard procedure format: instead of inputs, steps, and validation, dream opens space for emergence without evaluation. Use before design work, before naming things, before choosing between approaches, or anywhere premature structure kills the best option.
Evolve Skill From Traces
45pjt222/agent-almanac
Evolve SKILL.md files from agent execution traces using a three-stage pipeline: trajectory collection from observed runs, parallel multi-agent patch proposal for error and success analysis, and conflict-free consolidation of overlapping edits via prevalence-weighting. Based on the Trace2Skill methodology.
Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.