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 mimfort/rag_for_git --skill summarize-subsystemsgit clone --depth 1 https://github.com/mimfort/rag_for_gitWrote 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/mimfort/rag_for_git/summarize-subsystems)<a href="https://agentmods.dev/skills/mimfort/rag_for_git/summarize-subsystems"><img src="https://agentmods.dev/badge/skills/mimfort/rag_for_git/summarize-subsystems/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/mimfort/rag_for_git/summarize-subsystems"><img src="https://agentmods.dev/badge/skills/mimfort/rag_for_git/summarize-subsystems.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Prompt Injection · line 18 Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
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 | $0.00089 | $0.02795 |
| Opus 5.5 | $0.00036 | $0.01118 |
| Sonnet 5.5 | $0.00018 | $0.00559 |
| Haiku 4.5 | $0.00009 | $0.00280 |
Grade A, and why
summarize-subsystems 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 1mo 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 — 169 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Summarize subsystems (community summaries)
Cluster the base code graph into subsystems (by module path) and write a short, grounded
summary for each, persisted for ask / PR-walkthrough to use as a cheap high-level prior. This
skill reads code and writes summaries to the reviewer store; it does NOT modify code or post to
GitHub.
Always write summaries and answer the user in Russian (the project language), regardless of this
file's language. Tool calls, code identifiers and path:line stay verbatim.
Tools
Plus list_subsystem_clusters, get_subsystem_summary_work, get_file_skeletons,
index_subsystem_summary, prune_subsystem_summaries and backfill_summary_embeddings
(reviewer MCP). File-summary jobs read source ONLY through get_file_skeletons; the harness
Read is not used on source files in this skill.
Pipeline
- Resolve repo/branch.
-
List clusters (compact, paginated). Walk every page: call
list_subsystem_clusters(repo, branch, compact=True, limit=100), then repeat withoffset += 100while the response hashas_more == true. Stop only whenhas_more == false— a partial walk is not a full pass. Empty /noteabout an empty index → tell the user (in Russian) to runrag-reviewer:sync-codebasefirst, then stop.Every page carries the same whole-set fields:
depth(the applied cluster depth),layout_token(server-owned identity of the effective default depth plus sorted per-prefix overrides),depth_source(env|.review.yml|arg),deferred(stale clusters held back this pass under the cost cap, envSUMMARY_REBUILD_CAP),deferred_files(their pending file jobs),orphans(stored summaries whosecluster_keyis no longer a current cluster), andtotal_clusters(non-deferred clusters across the whole set, page-independent).The compact listing deliberately does not carry file-level data: each cluster gives
cluster_key,num_members,source_hash,stale,bootstrap,full_rebuild,reused_filesand the numeric countersadded,changed,removed,moved— no paths and no fingerprints. That keeps the listing O(clusters) instead of O(files). Per-cluster file detail comes fromget_subsystem_summary_workin step 5.
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.
- 1mo ago First seen · 169 lines · 89 tokens per session scan A 9b0677672cbc
summarize-subsystems is a skill published in the GitHub repository mimfort/rag_for_git (20 stars, last pushed 6d ago), licensed MIT. It adds 89 tokens to every session and 2,795 once invoked, about $0.0004 per session on Opus 5.5. 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
code-knowledge-graph
Codebase'i knowledge graph olarak analiz et. Dependency, call graph, hotspot analizi.
m1nd-first
Use when investigating a repository, searching for implementation, reviewing changes, working from specs/docs, or preparing a risky code change in an environment where m1nd is available. This doctrine makes m1nd the first investigative layer before grep, glob, or manual file reads, except when the task is pure…
coograph-review
Strict read-only review of the current changes for spec compliance, convention violations, logic bugs, and architecture. Every finding cites a verbatim quote from a fresh file read. Use after implementation, before declaring done.
graphenium
Use when analyzing codebase structure, verifying pre-flight policy compliance, tracing transitive dependency closures, or auditing post-edit scope creep.
celmis-search
Search and read code through Celmis instead of grepping or cat-ing whole repositories. Use when you need where a symbol is defined, who calls it, what a file or module contains, how another service does something ("make a DB connection like in service X", "do authorization the way service X does", "how does X load its…
autoreview
Structured code review through the shared OpenClaw agent-skills installation.