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.
git clone --depth 1 https://github.com/gkaria/vibe-learnWrote 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/commands/gkaria/vibe-learn/explain)<a href="https://agentmods.dev/commands/gkaria/vibe-learn/explain"><img src="https://agentmods.dev/badge/commands/gkaria/vibe-learn/explain/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/commands/gkaria/vibe-learn/explain"><img src="https://agentmods.dev/badge/commands/gkaria/vibe-learn/explain.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 | $0.00029 | $0.00907 |
| Opus 5 | $0.00015 | $0.00453 |
| Sonnet 5 | $0.00006 | $0.00181 |
| Haiku 4.5 | $0.00003 | $0.00091 |
Grade B, and why
explain scanned grade B with 1 finding 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
3. The install directory this project's hooks point at: read the vibe-learn hook command from `.claude/settings.local.json` (or `~/.claude/settings.json`) — it ends in `<install-dir>/scripts/bootstrap.sh`, and `knowledge How it starts
The opening of the file, as written. The whole thing — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Read .vibe-learn/session-log.jsonl to see what this session touched and why (the user_prompt events carry the intent). Read .vibe-learn/knowledge.json if it exists (the cross-session knowledge ledger).
Knowledge helper: to mark concepts as seen, use the knowledge.sh helper via Bash — never hand-edit the JSON. Locate it in this order:
vibe-learn-knowledgeon PATH (plugin install — check withcommand -v vibe-learn-knowledge)~/.vibe-learn/scripts/knowledge.sh(global install)- The install directory this project's hooks point at: read the vibe-learn hook command from
.claude/settings.local.json(or~/.claude/settings.json) — it ends in<install-dir>/scripts/bootstrap.sh, andknowledge.shsits in the same directory scripts/knowledge.shin the vibe-learn repo, if this project is the vibe-learn repo itself
If none of these exist, skip the ledger step silently — the tour itself still works.
Parse $ARGUMENTS to pick the target:
Target: none
If $ARGUMENTS is empty: tour the most significant file or subsystem this session touched. Prefer, in order: a newly created file that other touched files import; the file edited most often; the file behind the goal stated in the last user_prompt. Say which one you picked and why in one line.
Target: <file>
If $ARGUMENTS names a path that exists: tour that file.
Target: <topic>
Otherwise treat $ARGUMENTS as a topic (e.g. auth flow, the retry logic): find the files behind it — start from the session log, then grep the codebase — and tour them as one subsystem. If nothing matches, say so and list what the session did touch.
Tour structure (all targets)
Read the target file(s) and their immediate callers and callees before writing anything. Walk the code top-down as a colleague would at a whiteboard:
- Entry point — where execution starts and what triggers it (a request, a hook event, a CLI call, an import).
- The spine — the 3–5 load-bearing pieces, in execution order. For each, say why it is there, not just what it does. Quote the real line with a
path:linereference. - The edges — error paths, ordering constraints, defaults, and anything that would break if rearranged or removed.
- Connections — what calls this, what this calls, and where a change here would ripple.
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 · 57 lines · 29 tokens per session scan B edf72db63649
explain is a command published in the GitHub repository gkaria/vibe-learn (54 stars, last pushed yesterday), licensed MIT. It adds 29 tokens to every session and 907 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-07.
Other commands, from other repositories
help-flow
Help about Rosetta: explains capabilities and usage.
merge
Finish a PR properly: every check green, every review addressed — human and bot — then merge and clean up.
pr
Prepare and open a pull request the senior way: gate, template, scrubbed, everything visible.
spec
Spec-first design: a gap-closing interview that produces a complete spec, with a quality controller that blocks until every section is answered and every question resolved.
plan
Turn an approved spec into an implementation plan an engineer with zero context could execute — with a quality controller that blocks placeholders and hollow tasks.
todo
The quality-gated task list: tasks with real descriptions, testable acceptance criteria, and evidence — a task only closes when the controller agrees it is done.