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/kreek/consult/databasenpx skills add kreek/consult --skill databasegit clone --depth 1 https://github.com/kreek/consultWhat 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 | $0.00019 | $0.01368 |
| Opus 5 | $0.00010 | $0.00684 |
| Sonnet 5 | $0.00004 | $0.00274 |
| Haiku 4.5 | $0.00002 | $0.00137 |
Grade A, and why
database 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 yesterday.
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 — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Database
Iron Law
PROTECT PRODUCTION DATA FIRST: PROVE ROLLOUT, LOCKING, AND RECOVERY BEFORE CHANGE.
When to Use
- Schema design, migrations, indexes, query plans, isolation levels, connection pools, soft delete, N+1 fixes, online DDL, transactional outbox/CDC, or production data changes.
When NOT to Use
- API contract design; use
api. - Rollout sequencing outside the database; pair with
release. - Cache freshness and invalidation; use
performance.
Core Ideas
- Destructive and hard-to-reverse database changes are the user's call.
Route schema and migration changes that other code or stored data binds
to through
contract-first, which owns the approval scope; data deletion and non-reversible backfills always need explicit approval for data safety. An additive, easily reversed change on a development-stage schema proceeds with the shape stated in the close-out. - Use the project's existing database unless the task is choosing a store.
For greenfield defaults and store-selection caveats, use
architecture. - Expand, migrate, verify, switch, then contract in separate deployable steps.
- Review SQL and lock behavior, not just ORM code.
- Backfills are batched, resumable, observable, and reversible.
- Constraints enforce invariants. Every uniqueness invariant needs a DB-level
UNIQUE,EXCLUDE, composite, or partial equivalent. Application-layer checks race under concurrency. - Indexes and plans follow real access paths. New foreign keys and known
WHERE,JOIN, orORDER BYpredicates need supporting indexes in the same migration, or a stated reason they do not; query changes need EXPLAIN/ANALYZE plans on production-shaped data. - Isolation level is a design decision; retries are part of serializable correctness.
- State changes and durable publication need atomicity through transactional outbox, CDC, or an equivalent handoff when the two cannot silently diverge.
- Data recovery is part of the change: backup/PITR must cover the blast radius.
What ships with it
3 files 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.
- yesterday First seen · 120 lines · 19 tokens per session scan A fdd3bdf404f1
database is a skill published in the GitHub repository kreek/consult (1 stars, last pushed 6d ago), licensed MIT. It adds 19 tokens to every session and 1,368 once invoked, about $0.0001 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
argent-tv-interact
Control and inspect TV apps via argent — Apple TV (tvOS), Android TV (leanback), and Amazon Fire TV (Vega). Boot the target, read focus, navigate with the D-pad remote, type, screenshot, and on Vega debug the JS runtime (evaluate, console logs, network inspector). Use when a task targets a TV (runtimeKind "tv", or…
review
5-pass structured code review — correctness, security, performance, readability, consistency.
live-preview
Mid-build visual verification loop. Takes screenshots of components during construction, not just after. Catches visual regressions and invisible features before they compound. Requires Playwright or similar screenshot tool.
marshal
Meta-orchestrator that takes any direction — broad, specific, or vague — and autonomously chains skills and context into actionable work. Gathers context from codebase, docs, and memory. Only asks the user when it genuinely cannot proceed. Single-session orchestrator.
wiki
Markdown-first knowledge base where the LLM acts as librarian. Ingests raw sources, compiles and interlinks topic files, self-maintains an index. No vector DB or embeddings required -- uses LLM-native navigation over structured markdown up to 400K words.
review-offered-task
Review a task that has been offered to you and decide whether to accept or reject it.