dotfiles

Deep codebase analysis — architecture health, test coverage gaps, tech debt, and actionable findings

Run pre-commit checks (fmt, clippy, test) then stage, commit, and push with a conventional message

Review staged or recent changes for correctness, style, and Rust conventions before committing

Use when SSHing into remote machines, looking up credentials, renaming Tailscale devices, or accessing services on the tailnet. Symptoms - SSH auth failures, too many authentication failures, command not found tailscale, need to find credentials for remote machines.

Use when mixing Tokio async code with synchronous blocking I/O libraries (tar, flate2, zip, csv), or when seeing "cannot block the current thread from within the async context" panics, or needing to stream async bytes into a sync reader.

Structured BAML schema edit → validate → test loop for the devloop project. Use when editing any *.baml file in crates/baml/baml_src/ — covers version parity gotchas, uvx invocation, regeneration workflow, and the analyze verify step.

Use when reading large files (logs, JSONL, CSVs) that exceed context limits, when the Read tool returns truncated output, or when needing to locate a specific section in a multi-thousand-line file without reading everything.

Use when running daily maintenance across all repos — pull, analyze, fix P0s, synthesize to Obsidian, push. Invoke via /daily-orchestration or at start of day.

Use when running devloop council analysis on a repo. Covers the correct op run invocation, model selection, and narrative output format. Invoke instead of `just analyze` which is broken.

Full benchmark cycle for devloop — collect criterion results, check against budgets, commit the run to SQLite, surface regressions. Use after running cargo bench or when managing performance budgets.

Use when asked to update today's daily note, write a standup, or summarize what happened in a repo and record it in the Obsidian vault. Combines devloop council analysis with Obsidian daily note creation.

Use when asked to summarize recent repo activity, show what happened, or give a timeline view of work — with or without a time window argument

Use when converting a markdown task list, numbered list, or planning breakdown into doob todos. Symptoms - "add these tasks to doob", "import this plan", "create todos from this list", task breakdowns that need tracking.

Use at the start of any doob work session to load situational awareness — pending todos, overdue items, kanban state, and recent git activity. Use before starting new features, triaging work, or resuming after a break.

Use when implementing a new sync provider adapter for doob — creating a new IssueTracker trait implementation, wiring config, and writing tests. Symptoms - "add GitHub Issues sync", "implement Linear adapter", "new sync provider".

Use when assessing which roadmap phase doob is currently in, what's code-complete vs stub vs unstarted, or deciding what to work on next. Use before starting a new phase or when the README roadmap and codebase feel out of sync.

Use before any change to src/db/schema.rs or SurrealDB schema definitions. Symptoms - adding a new field to todo/note, changing an index, adding a new table, or seeing schema-related test failures after a schema edit.

Prioritized todo triage for the current project — runs doob todo list filtered to current repo, sorts by priority score, picks the highest-priority item to start, marks it in-progress, and creates a task checklist. Use at the start of a session or when asking "what should I work on next?"

Use when env vars are missing, wrong, or when debugging why `op run` or `direnv` isn't loading the right secrets. Traces the full source_up chain from project to root, showing which .envrc files are loaded and what each one injects.

Use when op run fails with env conflicts, direnv isn't loading secrets, source_up chain is broken, Claude shell can't resolve op:// URIs, or secrets work in terminal but not in Claude context

Use when synthesizing cross-project activity at end of session, generating a cross-repo narrative summary, or consolidating work from multiple repos into the Obsidian daily note

Use when setting up a new developer workstation for the Maestro project, when a developer reports missing tools or broken auth, or when onboarding someone to Maestro development

Use when working on minibox CI, managing the self-hosted GHA runner on $INFRA_VPS_HOST, running xtask gates, or diagnosing CI failures. Symptoms - CI failing, runner offline, mise not found, xtask error, need to SSH into $INFRA_VPS_HOST.

Use when developing minibox — quality gates, adding crates/adapters, testing strategy, VPS Linux testing, bench ops, and common workflows. Symptoms - need to run tests, add a new crate, wire an adapter, run benchmarks, deploy to VPS, or understand the development workflow.

Use when debugging runtime version issues, toolchain not found errors, or version conflicts across projects. Symptoms - wrong node/rust/python/go version active, mise shim errors, RUSTUP_TOOLCHAIN conflicts, tool not found after install, global version overriding project version.

Structured workflow for obfsck feature work — covers ObfuscationLevel logic, config/secrets.yaml group gating, the PII min_level invariant, and the TDD loop for adding patterns or flags. Use when implementing any obfsck feature from the backlog.

Use when working in the Obsidian Vault directory — creating notes, editing vault content, running vault scripts, or answering questions about projects/experiments/infrastructure tracked in the vault

Use before starting work on a feature or bug to get historical context from 9 months of LTM. Use when you need to know "what did I try before", "why was this decision made", or "what was the state of X last month".

Use when working with Pieces — the on-device AI memory and productivity platform for developers. Covers: PiecesOS setup and architecture, Long-Term Memory (LTM) queries, MCP integration with Claude/Cursor/GitHub Copilot/VS Code/Zed/JetBrains and 15+ other clients, CLI usage (save/search snippets), Desktop App (Timeline, summaries, Copilot), IDE plugins (VS Code, JetBrains), and the Obsidian plugin. Trigger on: "pieces", "PiecesOS", "LTM", "Pieces MCP", "pieces copilot", "pieces drive", "pieces obsidian", "pieces timeline", or any task involving Pieces integrations.

Use when working on any Rust crate or workspace — before implementing features, after editing .rs files, when adding or removing dependencies, or when CI is failing. Symptoms - clippy warnings treated as errors, fmt diff in CI, wrong edition error, workspace member not found, slow test runs.

Use when writing any standalone Rust script, one-off utility, or when tempted to reach for a shell script with complex logic. Covers rust-script shebang, inline Cargo manifest, and dependency declaration.

Review and accept/reject insta snapshot changes after cargo nextest runs. Handles the stale .snap.new footgun, shows inline diffs, and does safe bulk or selective acceptance. Use whenever cargo nextest produces .snap.new files.

Use when writing Rust tests that call set_var or remove_var, seeing "use of unsafe function" errors in Rust 2024 edition, or when tests pass individually but fail randomly in parallel CI runs due to environment variable races.

Use when managing encrypted secrets, extracting SSH keys, accessing remote machines via Tailscale, or working with op/sops/age. Symptoms - sops decrypt failed, age key missing, op CLI errors, SSH auth failures, too many authentication failures, tailscale device not found, need credentials for machines or services.

Use at end of a session (or on demand) to extract repeated tool patterns into a reusable skill. Triggers on "extract this as a skill", "save this as a skill", "I keep doing this", or when you notice the same command sequence run 2+ times.

Standardized tool set definitions for Claude Code agents ensuring consistent tool access across similar agent types

Use when needing to compute a side effect (hash, checksum, byte count, progress) on streaming bytes without buffering, or when a function accepts impl Read and you need to intercept the bytes passing through it.

Use after a significant commit or when CI fails to run the devloop → doob → devkit workflow pipeline. Conductor creates doob tasks from findings and summarizes health. Invoke via /conduct or let forge dispatch it after commits.

Use when needing development context — devloop visualizes git commit history and Claude AI session activity for any repo. Available globally via ~/.local/bin/devloop.

Primary dev companion for minibox, devloop, doob, and devkit. Use for design discussions, debugging, refactoring, prototyping, and anything that doesn't fit a specific workflow. Forge auto-dispatches to sentinel, navigator, or conductor as needed — no confirmation required.