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 svy04/ballast --skill researchergit clone --depth 1 https://github.com/svy04/ballastWrote 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/svy04/ballast/researcher)<a href="https://agentmods.dev/skills/svy04/ballast/researcher"><img src="https://agentmods.dev/badge/skills/svy04/ballast/researcher/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/svy04/ballast/researcher"><img src="https://agentmods.dev/badge/skills/svy04/ballast/researcher.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00057 | $0.00911 |
| Opus 5 | $0.00028 | $0.00456 |
| Sonnet 5 | $0.00011 | $0.00182 |
| Haiku 4.5 | $0.00006 | $0.00091 |
Grade A, and why
researcher 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 10d 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 — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Researcher — collection delegated, judgment kept
Two different jobs hide inside "research": going out to gather, and deciding what to believe. This skill delegates the first and forbids delegating the second. The failure mode this prevents: a second model's fluent summary walking straight into your files as fact.
Setup (optional, sibling of the verifier)
If the project has .claude/ballast.researcher.json (this skill reads it — the rules hook does not; a copy-ready example ships in the ballast repo at rules/ballast.researcher.example.json):
{ "command": "your-researcher-cli --search" }
run that command with the research question as its final argument. Any CLI that takes a question and returns findings works — a different vendor's model is fine here too.
The config is a bare object — no version/rules wrapper (that wrapper belongs to rule catalogs). A file that is missing, unreadable, or empty behaves the same way: you collect directly, as always. A file that parses but has no usable command (wrapped, wrong shape) gets the same fallback — plus one sentence telling the user their config isn't the expected shape. The user can also invoke this skill directly as /ballast:researcher.
Health check — once per session, out loud
Before the first delegation of a session, run the command once on a throwaway question (or --version if the CLI has one) — don't spend a real query to discover the tool is dead. If it fails or hangs, say so once — "researcher CLI isn't responding; collecting directly" — then proceed yourself.
A configured researcher that silently never ran is worse than none, and retrying a dead tool burns the session. One check, one sentence, move on.
Everything returned is hearsay
- The researcher's output arrives as
hearsay, with the researcher named as the source — never as fact, however fluent it reads. - Promotion to
confirmedgoes through the verify gate like any claim: refute first, primary sources you actually opened, sample, limits, date. hearsaylives in working notes and session logs —memory/knowledge/stays gate-passed only (the knowledge-base skill owns that door).- Delegation splits the work, not the responsibility. The collecting was outsourced; the believing was not.
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.
- 10d ago First seen · 51 lines · 57 tokens per session scan A 7f6069cd62e4
researcher is a skill published in the GitHub repository svy04/ballast (71 stars, last pushed 16d ago), licensed MIT. It adds 57 tokens to every session and 911 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 2026-08-30.
Other skills, from other repositories
alive:demo
Generate a believable, lived-in ALIVE world from a free-text persona description (custom path) or a deterministic sandbox preset. Routes the create/list/activate/deactivate/delete/status surface and orchestrates the 5-stage subagent generation pipeline.
alive:system-upgrade
Upgrade ALIVE to the current version. Handles v1/v2/v3.x source states, multi-surface aware (alive-mcp / Hermes / Codex), retroactive version detection, partial-failure resume, dry-run previews, and rollback inspection.
alive:session-history
Revive sessions (quick or heavy), browse, and search — 'what happened recently?', 'find the session where we discussed X', 'revive yesterday's session'. For single-session recall and multi-session browsing. If the human needs to merge multiple sessions into one working context or detect conflicts between parallel…
alive:world
The human doesn't know what to work on, or wants to see everything at once. They need the big picture — what's active, what's stale, what needs attention. Renders a live world view grouped by ALIVE domain, then routes to open, tidy, find, history, or map.
alive:mine-for-context
Deep context extraction from source material. Creates reference bundles, builds extraction plans, tracks what's been extracted, and discovers new targets — people, subjects, patterns, connections. The archaeologist that turns raw sources into structured knowledge. Can be invoked by alive:session-history for targeted…
alive:settings
The human wants to adjust how the system behaves — not what it contains, but how it feels. Voice, rhythm, preferences, walnut-level config, full permissions toggleable, named squirrel config, action logging toggle. The system adapting to them. Routes to preferences.yaml or walnut config.yaml depending on scope. For…