truth_forge

Perceive the plugin ecosystem through the six-phase knowledge aggregation methodology — structural reconnaissance, hub identification, sampling, content extraction, pattern synthesis, and feedback. Use when the user says "what plugins do I need", "audit my plugins", "what's missing", "analyze my setup", "plugin gaps", "what should I build next", or wants to understand the current state of their Cowork plugin ecosystem. Also triggers when any plugin produces output that could inform what plugin to build next. This skill is the TRUTH phase of THE LOOP — it perceives what IS before meaning can be extracted. It reads not just structure but scars, grammar, traces, and the living relationships between plugins that structural analysis alone misses. This is a trace-aware analysis that comprehends how past agents thought about the ecosystem.

Create complete Cowork plugins through HOLD:AGENT:HOLD architecture and THE FORGE PROCESS. Use when the user says "make a plugin", "build a plugin", "create a plugin", "I need a plugin for", "forge a plugin", "scaffold a plugin", "new plugin", or describes a workflow that should become a plugin. Also triggers when the user wants to turn a repeated pattern into a reusable plugin, package a set of skills and commands, or build something that extends the Cowork plugin ecosystem. Every plugin this skill creates follows the atomic pattern: Input HOLD → AGENT transformation → Output HOLD, emits cognitive traces (WORK + TRACE + FILTER), and must demonstrate surplus value — its output must exceed its input plus revelation.

The recursive loop (ALPHA:OMEGA) that drives plugin evolution through THE FURNACE metabolic cycle. Each plugin this engine creates feeds its outputs AND TRACES back through TRUTH:MEANING:CARE to determine what plugin to build next. Use when the user says "evolve", "what's next", "keep going", "build the next one", "recursive loop", "plugin chain", "what does the ecosystem need now", "molt", or when any plugin's output reveals a gap that a new plugin could fill. This is THE LOOP — it doesn't just create plugins, it creates the conditions for the next plugin to emerge from the outputs of the last one. Every cycle must produce surplus value (output exceeds input) and emit cognitive traces (WORK + TRACE + FILTER). The loop never ends; it molts.

Generate professional content including reports, blog posts, summaries, emails, documentation, and creative writing. Use when the user wants to create, draft, write, or compose any text-based content. Includes templates for common formats, tone guidance for different audiences, and quality criteria for professional output. Also use when reviewing, editing, or improving existing content.

Connect to data sources, explore schemas, write queries, and profile data quality. Use when the user wants to query a database, explore a data source, profile tables, or understand their data schema. Handles SQL dialects, API-based data sources, and file-based data. Includes query optimization and result validation patterns.

__SKILL_DESCRIPTION__

Orchestrate multi-step automated workflows with trigger detection, step execution, error handling, and result delivery. Use when the user wants to automate a process, chain actions together, handle conditional logic between steps, or build trigger-action pipelines. Includes retry patterns, validation checkpoints, and execution logging.

Persist the agent layer. Every plugin, skill, pipeline, or knowledge operation MUST emit a cognitive trace alongside its output. Use when building plugins, creating skills, running pipelines, analyzing codebases, synthesizing knowledge, or any operation where the HOW matters as much as the WHAT. Triggers on: "make a plugin", "build a skill", "create a pipeline", "trace this", "capture the thinking", "persist the reasoning", "what did you decide", "show your work", "how did you get there". The core innovation: stop deleting what the system already generates. HOLD:AGENT:HOLD — this skill captures the AGENT.