Borrowing it
Nothing to install: this file belongs to rafaelghif/antigravity-agents. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/rafaelghif/antigravity-agents/main/.agents/skills/semantic-graphing/SKILL.mdgit clone --depth 1 https://github.com/rafaelghif/antigravity-agentsWrote 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/rafaelghif/antigravity-agents/semantic-graphing)<a href="https://agentmods.dev/skills/rafaelghif/antigravity-agents/semantic-graphing"><img src="https://agentmods.dev/badge/skills/rafaelghif/antigravity-agents/semantic-graphing/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/rafaelghif/antigravity-agents/semantic-graphing"><img src="https://agentmods.dev/badge/skills/rafaelghif/antigravity-agents/semantic-graphing.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 | $0.00036 | $0.00932 |
| Opus 5 | $0.00018 | $0.00466 |
| Sonnet 5 | $0.00007 | $0.00186 |
| Haiku 4.5 | $0.00004 | $0.00093 |
Grade A, and why
semantic-graphing 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 — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Semantic Codebase Graphing & Blast Radius Protocol
Role: Principal Code Intelligence & AST Architecture Specialist.
Overview & Trigger Conditions
Activate this skill when exploring new codebases, planning non-trivial refactors, identifying high-risk Architectural God Nodes, mapping symbol caller graphs, or evaluating the blast radius of API/schema modifications.
Trigger Scenarios & Keywords:
- Codebase reconnaissance, blast radius calculation, caller dependency trees, God Node identification, AST symbol scanning.
- Keywords:
semantic graph,knowledge graph,blast radius,call graph,dependency tree,ast scan,pagerank.
Core Standards & Invariants
-
Epistemic Exploration over Blind Grepping:
- Never blindly inspect large files or dump directories when understanding architecture.
- Extract the AST symbol dependency graph to identify module boundaries, central hubs, and interface contracts.
-
Mandatory Blast Radius Calculation:
- Before modifying or deprecating any shared function, utility, base class, or model, calculate the transitive blast radius: Identify ALL upstream callers that could break.
- Refactorings with a large blast radius (> 10 dependent files) require an Architecture Decision Record (ADR) and phased rollout.
-
Architectural God Node & Hub Awareness:
- Use PageRank centrality to identify core system hubs.
- High-centrality hub nodes require strict backward compatibility, zero breaking changes, and 100% test coverage.
-
Targeted Reading & Context Economy:
- Only read the specific line ranges identified in the semantic graph using
view_fileslice notation (StartLine/EndLine). Avoid reading irrelevant lines.
- Only read the specific line ranges identified in the semantic graph using
Graph Engine Commands (scripts/semantic_grapher.py)
- Full Codebase AST Scan & Hub Identification:
python3 scripts/semantic_grapher.py <dir> - Transitive Blast Radius Resolution:
python3 scripts/semantic_grapher.py <dir> --blast-radius <Symbol>Performs transitive BFS to detect all direct and indirect upstream dependents of a symbol. - PageRank Centrality Ranking:
python3 scripts/semantic_grapher.py <dir> --pagerank --top-central 10Ranks top 10 architectural hub symbols in the repository. - BFS Shortest-Path Dependency Tracer:
python3 scripts/semantic_grapher.py <dir> --path-find <StartSymbol> <EndSymbol>Traces end-to-end invocation flow between two disparate components. - GraphRAG Knowledge Graph Export:
python3 scripts/semantic_grapher.py <dir> --jsonExports structured JSON graph (nodes and edges) for ingestion.
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 Changed 9f9c2381f97b
- 3d ago Changed · +48 lines · +6 tokens per session d044e6f2e04b
- 4d ago Changed · +1 lines 76d05bcfd60e
- 9d ago First seen · 29 lines · 30 tokens per session scan A e0aa158adaa4
semantic-graphing is a skill published in the GitHub repository rafaelghif/antigravity-agents (5 stars, last pushed 2d ago), licensed MIT. It adds 36 tokens to every session and 932 once invoked, about $0.0002 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
coordinate-recover
Diagnose and safely resume Coordinate Agents Tasks after executable failure, non-zero exit, timeout, stale claim, processing message, or Implementer ERROR. Recovery is explicit and never an automatic retry loop.
debug-protocol
Evidence-first debugging in 5 stages — REPRODUCE, EVIDENCE, DIAGNOSE, FIX, PROVE. Never touch code before the root cause is stated with evidence; fix the source, not the symptom; hand dead-end investigations to the root-cause-debugger agent. Use for any bug, error, crash, regression, unexpected behavior, or /toh-fix…
engineer-harness
Shared delivery harness for the main commands (/toh, /toh-plan, /toh-fix, /toh-vibe) — senior-engineer tool selection rules (search real docs, reproduce before diagnosing, verify by running), THE EVIDENCE RULE (only a quoted passing run counts as done), non-developer communication mode, and stage-aware next actions…
error-handling
Intelligent error handling — silently auto-fix routine errors (missing imports, undefined variables, type mismatches, missing dependencies, syntax and lint issues) and translate everything else into friendly, non-technical language with clear next steps. The user never sees raw error messages or stack traces. Use…
gha
Analyze GitHub Actions failures and identify root causes.
systematic-debugging
Methodical problem-solving approach for diagnosing and resolving code issues in any stack before attempting a fix.