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/vunm-io/passdown/passdown-intakenpx skills add vunm-io/passdown --skill passdown-intakegit clone --depth 1 https://github.com/vunm-io/passdownWrote 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/vunm-io/passdown/passdown-intake)<a href="https://agentmods.dev/skills/vunm-io/passdown/passdown-intake"><img src="https://agentmods.dev/badge/skills/vunm-io/passdown/passdown-intake.svg" alt="Measured on agentmods" 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 | $0.00056 | $0.01135 |
| Opus 5 | $0.00028 | $0.00567 |
| Sonnet 5 | $0.00011 | $0.00227 |
| Haiku 4.5 | $0.00006 | $0.00113 |
Grade A, and why
passdown-intake 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 6d 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 — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Passdown Intake
Turn raw captured notes into properly planned work. Weak capture tools (chat apps, quick notes) only need to drop a markdown file into an inbox; this skill is where a capable agent applies judgment.
Configuration (read first)
Build the effective passdown configuration root-to-nearest:
- Start with every applicable
AGENTS.mdfrom the workspace/repository root down to the current directory. - If a thin
AGENTS.mdexplicitly points to a parent file outside the Git root, read that parent too. - Merge
## passdownkeys from root to nearest; a nearer value overrides only the same key and inherits all omitted keys. - Resolve relative config paths against the directory containing the
AGENTS.mdthat declared that value, not against an arbitrary current directory.
The effective configuration should contain:
## passdown
- inbox: <path to inbox directory>
- planning: markdown | openspec
- plan_dir: <path for markdown plans> # required when planning: markdown
- targets: <repo1>, <repo2>, ... # known target repos for routing
If a required key is still missing, ask the user for that value and suggest
adding it to the appropriate AGENTS.md.
Inbox note contract
Notes are markdown files with minimal frontmatter. Never demand more structure from capture tools than this:
---
status: new # new | processed | discarded
target: unknown # target repo if the author could guess it
---
# <idea / thing to do>
<free-form content>
A malformed note is still a valid note — it is raw material, not an artifact.
Process
- Scan the inbox for notes with
status: new(treat missing frontmatter asnew). List them briefly for the user. - For each note, in order:
- Read it fully. Cross-check existing knowledge/specs in the workspace if relevant.
- If intent is unclear, ask the user — one question at a time.
- Before writing, resolve the inbox, target repo, planning directory, and schema paths. Verify that the current host has read access to inputs and write access to every intended output. Cross-repo work often needs the parent workspace opened as a workspace root or an explicit permission grant.
- If sandbox, permission, or network policy blocks a required path, stop and
report the blocked path and operation. Do not redirect
HOME, create substitute caches, or edit project/tool configuration as a workaround. - Decide the disposition:
- Actionable work → determine the target repo (use
target:hint, the routing rules in AGENTS.md, or ask). Create the planning artifact in the target repo using the workspace's planning convention. Forplanning: markdown, create a plan file in the target repo'splan_dirfollowing the structure of passdown'stemplates/plan.md: every task carries a[dispatch: external-ok]or[dispatch: main]tag plus Paths, Done criteria, and Verification. Forplanning: openspec, use the portableopenspec new changeCLI path; a host-specific command such as/opsx:proposemay be used only when that host actually provides it. Either way the artifact must be self-contained: an executor must be able to work from it without reading this conversation. - Knowledge, not work → refile into the workspace's knowledge location per its conventions.
- Obsolete/noise → mark
status: discardedafter confirming with the user.
- Actionable work → determine the target repo (use
- Close the loop: update the note's frontmatter to
status: processedand append a link to the created change/refiled location. The note stays where it is — it becomes the permanent record of where the idea came from.
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.
- 6d ago First seen · 107 lines · 56 tokens per session scan A 7b8cb9f5e824
passdown-intake is a skill published in the GitHub repository vunm-io/passdown (8 stars, last pushed 16d ago), licensed MIT. It adds 56 tokens to every session and 1,135 once invoked, about $0.0003 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
ai-loop
Runs a bounded spec-build-review development loop with explicit scope, stop conditions, and human approval gates for risky or ambiguous work.
architecture-decision-records
Comprehensive patterns for creating, maintaining, and managing Architecture Decision Records (ADRs) that capture the context and rationale behind significant technical decisions.
agent-harness-fault-injection
Use when an agent workflow needs deterministic recovery evidence for sandbox, MCP/tool, worker, checkpoint, memory, or orchestration failures.
aria
Designs the data model, API contracts, and structural foundation of the system.
peon-ping-log
Log exercise reps for the Peon Trainer. Use when user says they did pushups, squats, or wants to log reps. Examples - "/peon-ping-log 25 pushups", "/peon-ping-log 30 squats", "log 50 pushups".
conductor-implement
Executes the tasks defined in the specified track's plan. Use this to start or continue working on a feature, bug fix, or chore.