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.
npx agentmods add skills/jpicklyk/task-orchestrator/quick-startnpx skills add jpicklyk/task-orchestrator --skill quick-startgit clone --depth 1 https://github.com/jpicklyk/task-orchestratorWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00082 | $0.04016 |
| Opus 5 | $0.00041 | $0.02008 |
| Sonnet 5 | $0.00016 | $0.00803 |
| Haiku 4.5 | $0.00008 | $0.00402 |
Grade A, and why
quick-start scanned grade A with 0 findings 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 310 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Quick Start — MCP Task Orchestrator
Interactive onboarding that teaches by doing. Detects your workspace state and adapts.
Step 1: Detect Workspace State
Resolve the project rootId first: check session context for a rootId injected by the SessionStart hook, or read .taskorchestrator/config.yaml's top-level project.rootId (a file read, not an MCP call).
Call the health check to determine which path to follow:
get_context()
When a rootId is known, pass it to scope the check to this project: get_context(ancestorId="<rootId>"). When no rootId is known — the common case for a truly fresh workspace, or one that hasn't been bootstrapped yet — call unscoped exactly as shown.
If no active or stalled items exist — follow the Fresh-Start Path (Steps 2-8). If active items exist — follow the Orientation Path (Steps A-C).
Step 1.5: Project Anchor Bootstrap (if needed)
Before following either path, check whether this workspace has a project anchor yet:
- If
.taskorchestrator/config.yamldoes not exist at all, skip this step — that's the truly fresh workspace covered by the Fresh-Start Path below. Bootstrap can happen on a later run once a config file exists (e.g., after/manage-schemascreates one). - If
.taskorchestrator/config.yamlexists and already has a top-levelproject:block, read itsrootIdand use it for scoping throughout this session — no bootstrap needed. - If
.taskorchestrator/config.yamlexists but has noproject:block, offer to create one viaAskUserQuestion: "This workspace doesn't have a project anchor yet — want me to create one? It lets/work-summary,/create-item, and other skills scope to just this project if multiple projects ever share the same database."
If the user accepts:
- Determine a project name — from
$ARGUMENTS, conversation context, or by asking. - Create the anchor item at depth 0:
manage_items(operation="create", items=[{title: "<project name>", type: "project", priority: "low"}]) - Write the canonical block into
.taskorchestrator/config.yaml:project: rootId: "<created-item-uuid>" name: "<project name>" - Older servers may not expose it, so check the tool list before calling — if a
manage_project_configtool is available, push the full current file text (not just theproject:block — the server never reads that block itself; seereferences/config-format.md→ Project Scoping) so per-root schema resolution picks it up immediately without waiting on a config reload:manage_project_config(operation="push", rootId="<created-item-uuid>", configYaml="<full current file text from step 3>")- Success → the returned
fingerprintconfirms the push landed; re-pushing identical content later returns the same fingerprint (idempotent). VALIDATION_ERROR→ surface the parse error to the user; the config.yaml write from step 3 is already saved locally, so nothing is lost — tell them to fix the file and retry the push (or run/manage-schemas validate).CONFLICT_ERROR(superseded) → the local file is older than the server's stored config (rare during onboarding — usually means another checkout already synced a newer version). Fetch the server's copy withmanage_project_config(operation="get", ...)and reconcile, or passforce: trueif overwriting is intentional.- A
warningfield → relay it to the user (non-fatal).
- Success → the returned
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.
- 3d ago First seen · 310 lines · 82 tokens per session scan A a7e7c9d1ce23
quick-start is a skill published in the GitHub repository jpicklyk/task-orchestrator (205 stars, last pushed 29d ago), licensed MIT. It adds 82 tokens to every session and 4,016 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
mulch-record-from-evidence
Turn the evidence of a finished work session — git commits, changed files, recently-touched seeds issues — into well-formed ml record invocations. Use at session close, when an agent has made changes worth preserving as mulch expertise but hasn't yet recorded them.
pneuma-session
Rewrite the active Pneuma session's UI title + one-line summary so the launcher and ProjectPanel rows reflect what the session is actually about. Use this skill whenever the user asks to "整理 / 概括 / refresh / re-title / summarize this session", whenever the conversation has produced substantive work and the default…
vc-context-discovery
Discover and load all relevant context for the current task. Lists feature group nested files with full paths, loads process/context/ files by domain routing. Called at the start of every agent session.
vc-test-coverage-plan
Use when creating a test plan for a blast radius. Assigns all 4 tiers (fully-automated, hybrid, agent-probe, known-gap) with exact commands, what each proves, and gap resolution options.
vc-predict
5 expert personas debate proposed changes before implementation. Catches architectural, security, performance, and UX issues early. Use before major features or risky changes.
vc-autopilot
Emit and validate the provisional goal block for Autopilot Mode. Owns the 9-field format and resume detection from a pasted goal block.