reskill

## Context When the coordinator hears "team, reskill" (or similar: "optimize context", "slim down charters"), trigger a team-wide optimization pass. The goal: reduce per-agent context consumption by extracting shared patterns from charters and histories into reusable skills. This is a periodic maintenance activity. Run whenever charter/history bloat is suspected. ## Process ### Step 1: Audit Read all agent charters and histories. Measure byte sizes. Identify: - **Boilerplate** — sections repeated across ≥3 charters with <10% variation (collaboration, model, boundaries template) - **Shared knowledge** — domain knowledge duplicated in 2+ charters (incident postmortems, technical patterns) - **Mature learnings** — history entries appearing 3+ times across agents that should be promoted to skills ### Step 2: Extract For each identified pattern: 1. Create or update a skill at `.squad/skills/{skill-name}/SKILL.md` 2. Follow the skill template format (frontmatter + Context + Patterns + Examples + Anti-Patterns) 3. Set confidence: low (first observation), medium (2+ agents), high (team-wide) ### Step 3: Trim **Charters** — target ≤1.5KB per agent: - Remove Collaboration section entirely (spawn prompt + agent-collaboration skill covers it) - Remove Voice section (tagline blockquote at top of charter already captures it) - Trim Model section to single line: `Preferred: {model}` - Remove "When I'm unsure" boilerplate from Boundaries - Remove domain knowledge now covered by a skill — add skill reference comment if helpful - Keep: Identity, What I Own, unique How I Work patterns, Boundaries (domain list only) **Histories** — target ≤8KB per agent: - Apply history-hygiene skill to any history >12KB - Promote recurring patterns (3+ occurrences across agents) to skills - Summarize old entries into `## Core Context` section - Remove session-specific metadata (dates, branch names, requester names) ### Step 4: Report Output a savings table: | Agent | Charter Before | Charter After | History Before | History After | Saved | |-------|---------------|---------------|----------------|---------------|-------| Include totals and percentage reduction. ## Patterns ### Minimal Charter Template (target format after reskill) ``` # {Name} — {Role} > {Tagline — one sentence capturing voice and philosophy} ## Identity - **Name:** {Name} - **Role:** {Role} - **Expertise:** {comma-separated list} ## What I Own - {bullet list of owned artifacts/domains} ## How I Work - {unique patterns and principles — NOT boilerplate} ## Boundaries **I handle:** {domain list} **I don't handle:** {explicit exclusions} ## Model Preferred: {model} ``` ### Skill Extraction Threshold - **1 charter** → leave in charter (unique to that agent) - **2 charters** → consider extracting if >500 bytes of overlap - **3+ charters** → always extract to a shared skill ## Anti-Patterns - Don't delete unique per-agent identity or domain-specific knowledge - Don't create skills for content only one agent uses - Don't merge unrelated patterns into a single mega-skill - Don't remove Model preference line (coordinator needs it for model selection) - Don't touch `.squad/decisions.md` during reskill - Don't remove the tagline blockquote — it's the charter's soul in one line --- > Source: [candoumbe/DataFilters](https://github.com/candoumbe/DataFilters) — distributed by [TomeVault](https://tomevault.io). <!-- tomevault:4.0:skill_md:2026-07-19 -->