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 skills add vincentmumme/personalos-boilerplate --skill task-managergit clone --depth 1 https://github.com/vincentmumme/personalos-boilerplateWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/vincentmumme/personalos-boilerplate/task-manager)<a href="https://agentmods.dev/skills/vincentmumme/personalos-boilerplate/task-manager"><img src="https://agentmods.dev/badge/skills/vincentmumme/personalos-boilerplate/task-manager/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/vincentmumme/personalos-boilerplate/task-manager"><img src="https://agentmods.dev/badge/skills/vincentmumme/personalos-boilerplate/task-manager.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00090 | $0.02459 |
| Opus 5 | $0.00045 | $0.01229 |
| Sonnet 5 | $0.00018 | $0.00492 |
| Haiku 4.5 | $0.00009 | $0.00246 |
Grade A, and why
task-manager 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 9d 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 — 147 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: task-manager
Contract
This skill guarantees:
- Each confirmed open commitment is one canonical
actionrecord underoperations/actions/<uuid>.md. - Each future reassessment without a current commitment is one canonical
attention-triggerunderoperations/attention-triggers/<uuid>.md. operations/todo.md, indexes, briefings, Daily records, project files and automation reports are never active task stores. They may only point to or derive views from the atomic records.- The exact schemas, sections, lifecycles and field rules come from
system/data-model/profiles/action.yaml,system/data-model/profiles/attention-trigger.yaml,system/templates/action.mdandsystem/templates/attention-trigger.md. - A source/import run reconciles matching existing records before creating anything. The system is not append-only.
- Only a real, confirmed, personally relevant open commitment becomes an Action. Inferred work, possible improvements, raw thoughts and ideas do not enter the active action set.
- A meeting discussion, opportunity, possible pilot, suggested preparation path, or hoped-for follow-up is not a confirmed commitment. Do not set
ready,action_priority,due, ortargetunless the source or {{user_name}} explicitly commits to the work and, for dates, to that exact timing. If no date was agreed, omit timing fields instead of manufacturing urgency. - A later date is not automatically a task. Use an Attention Trigger when the only present need is to reassess the situation later.
- A routine incoming receipt or subscription invoice is Finance collection truth, not an immediate Action. Preserve it in the proper Finance owner and add it to the month-end booking corpus. If an Operations record is needed for the month, maintain one
deferredbooking-batch Action whosenot_beforeis the first of the final five calendar days; do not create or surface one Ready Action per receipt. Real payment deadlines, dunning, suspension, cashflow, legal, disputed-performance, tax, or payment-decision risks remain separate Actions. waitingis an Action lifecycle because the commitment already exists; it requireswaiting_onandfollow_up_at.- Project and domain files own durable outcome truth. Action records own the current execution boundary, timing and done condition. No mirrored checkbox list is allowed.
- When source evidence changes both durable truth and an Action, update both owners in one bounded mutation.
- Before closing an Action, propagate the result to its durable owner. Then set the Action to
completedorcancelled; do not erase material audit history merely to shorten a list. - Every Action has an observable
Done Boundary, exactly one currentNext Action, evidence links and one execution mode:agent-execute,agent-prepareorhuman-only. - Keep relations typed: sources and receipts belong in
evidence_refs; affected people, companies, projects and domain owners belong in optionalaffected_owner_refs. Never use an owner record as evidence merely to satisfy the required field. - Dates have separate meanings:
dueis a real deadline,targeta planned finish,not_beforehides deferred work until a date, andfollow_up_atrechecks waiting work. - Prioritization is a derived, time-sensitive view. Do not encode old
P0/P1/P2buckets as permanent record truth unless the current profile explicitly admits anaction_priorityvalue. - Google Tasks is not read, reconciled or mutated by PersonalOS operations.
- Every mutation follows [[system/contracts/core/personalos-mutation-contract]] and [[system/runbooks/core/personalos-mutation]] and ends with
pos-verifyon the explicit changed-file set. There is no general write-skill between this owning capability and the target records. - After a material Action or Attention Trigger state change, invoke [[skills/priority-dashboard/SKILL]] with
--reason operational-state-change. This is an explicit post-mutation rebuild, not a realtime watcher and not a second lifecycle write.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 9d ago First seen · 147 lines · 90 tokens per session scan A 83779ed3d9ac
task-manager is a skill published in the GitHub repository vincentmumme/personalos-boilerplate (17 stars, last pushed 14d ago), licensed MIT. It adds 90 tokens to every session and 2,459 once invoked, about $0.0005 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 2026-08-31.
Other skills, from other repositories
wrap-session-up
End-of-session review for a Claude Code conversation. Replays what was discussed, checks for open action items, documents relevant outcomes in the brain vault, and creates Todoist tasks for anything untracked. Use at the end of a work session to ensure nothing falls through the cracks.
gws-obsidian-prep
Cross-skill recipe: fetch today's (or tomorrow's) Google Calendar events via gws CLI and create meeting prep notes in the Obsidian vault.
resume
A workflow for reopening a paused task by scanning its saved project notes, blockers, logs, and delegated work. It presents the parked tasks so the user can choose which one to continue.
vault-onboarding-guide
An interactive guide for learning a file-based knowledge store with an agent. It teaches how to organize incoming material, projects, plans, tasks, logs, parked work, and resumed work.
obsidian
Obsidian vault management — create notes, backlinks, tags, and Dataview queries via file operations.
weekly
Run the weekly review of the big-brain second brain: triage the inbox, walk active projects, reschedule or drop overdue tasks, check vault health, and write a review summary. Use when the user types /weekly, or asks for a "weekly review", "review my projects", "clean up my brain", "triage my inbox".