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/rilldata/rill/reflectnpx skills add rilldata/rill --skill reflectgit clone --depth 1 https://github.com/rilldata/rillWhat 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 | $0.00011 | $0.00181 |
| Opus 5 | $0.00005 | $0.00090 |
| Sonnet 5 | $0.00002 | $0.00036 |
| Haiku 4.5 | $0.00001 | $0.00018 |
Grade A, and why
reflect 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.
What it actually says
Extract learnings from this conversation and persist them to the appropriate project files.
Focus area: $ARGUMENTS
Instructions
- Review the conversation for patterns, conventions, or decisions worth persisting
- Categorize each learning by where it belongs:
CLAUDE.md— General conventions, project guidance, backend patterns.claude/rules/*.md— Scoped rules (frontend, docs, code review, etc.).claude/skills/*/SKILL.md— Improvements to skills
- Draft proposed changes and present them via
AskUserQuestionfor approval before applying
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 · 19 lines · 11 tokens per session scan A bad314120f66
reflect is a skill published in the GitHub repository rilldata/rill (2,853 stars, last pushed 3d ago), licensed Apache-2.0. It adds 11 tokens to every session and 181 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 2026-08-30.
Other skills, from other repositories
hardwood-pr-git
Rebase, rewrite, squash, reword and push Hardwood PR history — including contributor PRs whose author and committer identity must survive the rewrite. Use whenever the user asks to "rebase", "rebase onto main", "fix the commit message", "prefix with the issue key", "squash the commits", "force push (with lease)"…
hardwood-cli
Inspect, debug, and convert Apache Parquet files with the hardwood CLI. Use whenever a user is debugging code that reads or writes Parquet (Arrow / Spark / Flink / Java / Python) and the data file itself may be the cause — to check the schema, physical and logical types, repetition, or a row sample. Use it to diagnose…
hardwood-review
Review a Hardwood PR, local branch, or staged diff against the project's CLAUDE.md conventions and recurring failure modes. Use whenever the user asks to "review", "look over", "check", "audit", or "assess" a PR number, a branch, or pending changes in the hardwood-hq/hardwood repo. Also use when the user pastes a PR…
hardwood-address-review
Work through a reviews/pr- -review.md findings file against a Hardwood PR — check out the PR branch, make the fixes, surface each Decision with its options + recommendation for the maintainer to pick, tick off addressed items in the review file, and commit (do not push). Use whenever the user says "address review for.
parquet-dev-ml
Search the Apache parquet-dev mailing list by downloading the raw monthly mbox archives from lists.apache.org and grepping them locally. Use whenever the user wants to find, read, or quote a thread on the Parquet dev list ([email protected]) — e.g. "search the parquet-dev ML", "find the thread about X on the…
exasol-import-export
Exasol IMPORT and EXPORT SQL statements: syntax, file formats (CSV, FBV, Parquet), cloud storage (S3, Azure, GCS), connection objects, error handling, and ETL staging patterns.