kb-onboard

A session-starting guide that reads the latest handoff note and project overview from an Obsidian knowledge base, a folder of linked notes. It gives the coding agent the project's current context.

In plain words
What is it for?
Use it when starting work on a project with an Obsidian vault, or when you need to continue previous work and catch up on the project state.
Why use it?
It reduces the time spent reconstructing what happened in an earlier work session and helps prevent missed instructions or duplicated work.

Skill for Claude CodeCodex

Part of the obsidian-kb plugin — 8 skills, 8 commands shipped together

Install

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.

agentmods
npx agentmods add skills/waelmas/codeplow/kb-onboard
Any agent
npx skills add waelmas/codeplow --skill kb-onboard
Clone the repo
git clone --depth 1 https://github.com/waelmas/codeplow

Made for: Claude Code, Codex.

Or install obsidian-kb, the plugin that ships this one along with the rest of its 8 skills, 8 commands.

Per session 98 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,173 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00098 $0.03173
Opus 5 $0.00049 $0.01587
Sonnet 5 $0.00020 $0.00635
Haiku 4.5 $0.00010 $0.00317

Measured 3d ago against content hash a6adc2c39302, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

kb-onboard 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.

Tells the agent never to refusehighAnti-refusal

Suppressing the ability to decline removes a core safety control; a later harmful request then succeeds.

- **Don't refuse.** If you can't find a perfect handoff, work with what's available. Something is always better than nothing.
obsidian-kb/skills/kb-onboard/SKILL.md · 240 lines

How it starts

The opening of the file, as written. The whole thing — 240 lines — stays where its author put it; the contents beside it link to each section on GitHub.

kb-onboard - Session Briefing

Read the latest session handoff and project overview to quickly brief yourself (or the agent) on the current state of the project. Lean and fast - get context, then get to work.

Robustness rules (read first)

These rules exist because non-interactive invocations (claude -p "...", Copilot CLI -p, etc.) can end the agent's turn prematurely in the middle of multi-step skills. Rule #6 is the performance rule that matters most on smaller models.

  1. Run every Bash call in this skill in the FOREGROUND. Never set run_in_background: true on any command here. Background completions can be misread as the end of your turn, causing the skill to terminate after preflight. If you need parallelism (e.g., listing + searching), call the commands sequentially instead.

  2. Complete Steps 0–5 before handing back control. The briefing in Step 4 is NOT an end-of-turn signal. Always produce it, then continue.

  3. If the user's message contained instructions beyond /kb-onboard, e.g. "/kb-onboard, then answer these questions" or "/kb-onboard and continue the feature" - complete the briefing first, then proceed with those instructions in the SAME turn. Do not stop after the briefing.

  4. If CLI calls (obsidian vaults, obsidian vault info=name, etc.) fail or return unexpected values, fall through to filesystem mode immediately (CLI_MODE=0). A contested active-vault (from parallel agents or running scripts) is not a reason to abort - it's a reason to switch to find/cat/grep on $VAULT_PATH and keep going.

  5. Emit a deterministic completion marker at the end of Step 5: literally output the line <!-- kb-onboard:complete --> on its own line as an HTML comment. This tells both humans and downstream tooling that the skill ran to completion, and lets you confidently move on to any remaining instructions.

  6. FAST PATH when the user provides an explicit vault PATH. If the user's message specifies a filesystem path to the vault (e.g., "The vault is at /path/to/Foo KB" or "use the vault at ~/vaults/Foo"), you have everything you need to skip almost all of this skill. Concretely:

    • Set VAULT_PATH=<provided path>, CLI_MODE=0.
    • Skip Step 0 (preflight), Step 1 (resolution algorithm), Step 1b (Vault Access Sequence) entirely. Do NOT run command -v obsidian, obsidian vaults, or obsidian vault info=name.
    • Read exactly TWO files and stop: the newest file in $VAULT_PATH/Sessions/ (use a single ls -t Sessions/ | head -1 to pick it), and $VAULT_PATH/Index.md (if present; if not, skip). Do NOT enumerate candidate overview filenames. Do NOT run obsidian search. Do NOT read any Architecture/ or other vault subdirectories unless the user explicitly asks a follow-up about them.
    • Emit a one-paragraph briefing (≤ 80 words) capturing TL;DR + Next Steps + one Watch-Out. Do NOT produce the full block-quoted briefing format from Step 4.
    • Emit <!-- kb-onboard:complete --> and continue to any remaining instructions.

Read the full file on GitHub · 240 lines

Changes

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.

  1. 3d ago First seen · 240 lines · 98 tokens per session scan C a6adc2c39302

Subscribe to this mod's changes

kb-onboard is a skill published in the GitHub repository waelmas/codeplow (8 stars, last pushed 4mo ago), licensed MIT. It adds 98 tokens to every session and 3,173 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it C with 1 finding (tells the agent never to refuse). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

auto-perf-optimize

Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.

microsoft/vscode · 62 tokens

chat-perf

Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.

microsoft/vscode · 51 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens