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/potpie-ai/potpie/potpie-source-ingestionnpx skills add potpie-ai/potpie --skill potpie-source-ingestiongit clone --depth 1 https://github.com/potpie-ai/potpieWhat 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 | $0.00082 | $0.02719 |
| Opus 5 | $0.00041 | $0.01359 |
| Sonnet 5 | $0.00016 | $0.00544 |
| Haiku 4.5 | $0.00008 | $0.00272 |
Grade A, and why
potpie-source-ingestion 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 — 249 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Potpie Source Ingestion
Use this skill for explicit ingestion requests. The harness is the intelligence: it gathers source data, reads it, decides what is durable, resolves identity, and writes semantic graph mutations with evidence. Potpie validates and stores; it does not decide what source material means.
Non-Negotiables
- Use a todo/checklist for every repository or multi-source ingestion. Do not jump directly from a README to graph writes.
- Local inspection is required for repo understanding. Scanner-driven graph
updates are forbidden. Inspect files with
rg,rg --files,git, and structured tooling; do not run legacy or deterministic ingestion/scanner commands that walk the tree and write graph facts. - Use subagents only for read-only discovery slices. The main agent owns source selection, identity resolution, mutation proposals, commits, and final synthesis.
- Do not write until each required discovery lane is complete, explicitly unavailable, or intentionally scoped out by the user.
- Every write needs source refs, source authority, truth class, confidence, compact summary, and retrieval-grade description.
Phase 0: Scope And Preflight
- Define source kind, pot/project, repo/path/URL, time window, and target memory shape: baseline, history, docs, infra, debug memory, preferences, or all.
- Verify Potpie scope and graph availability:
potpie --json pot info
potpie --json source list
potpie --json graph status
potpie --json graph catalog --task "harness-led source ingestion"
- If the repo is not registered, register metadata only. Use explicit
--potwhen pot scope is ambiguous:
potpie source add repo . --pot <pot-id-or-name>
- Describe the views you expect to write/read before authoring mutations:
potpie --json graph describe features --view feature_context --examples
potpie --json graph describe infra_topology --view service_neighborhood --examples
potpie --json graph describe recent_changes --view timeline --examples
potpie --json graph describe decisions --view preferences_for_scope --examples
potpie --json graph describe debugging --view prior_occurrences --examples
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 · 249 lines · 82 tokens per session scan A acf24c5e5449
potpie-source-ingestion is a skill published in the GitHub repository potpie-ai/potpie (5,702 stars, last pushed 2d ago), licensed Apache-2.0. It adds 82 tokens to every session and 2,719 once invoked, about $0.0004 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
build-paths-advanced
Scaffold an agent system where Oracle AI DB is the only state store, composed from the build-paths/skills/ building-block library. Stack — langchain-oracledb + oracle-database-mcp-server + in-DB ONNX embeddings + OCI GenAI Grok 4 + Open WebUI. Three projects — production-feeling NL2SQL+RAG hybrid analyst…
soccer-workshop-setup
Bootstrap the soccer analytics agent workshop. Starts the Oracle AI Database Free container, applies schema, loads the FIFA dataset, optionally trains models, populates LangChain OracleVS hybrid retrieval plus semantic memory, applies LangGraph OracleDB observability, and verifies OCI GenAI access. Use when starting…
design-taste-frontend
Senior UI/UX Engineer. Architect digital interfaces overriding default LLM biases. Enforces metric-based rules, strict component architecture, CSS hardware acceleration, and balanced design engineering.
skill-creator
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
build-paths-intermediate
Scaffold a Grok-4 tool-calling agent over an Oracle schema using langchain-oracledb + oracle-database-mcp-server + in-DB ONNX embeddings (registered MiniLM model, no external embedding API) + Open WebUI. For users who've built RAG before and want to rebuild it on the production-feeling Oracle stack.
oracle-mcp-server-helper
Wire the oracle-database-mcp-server into a Python project so an LLM agent can call listtables / describetable / runsql / vectorsearch at inference time. Handles install, stdio launch, and LangChain tool conversion via langchain-mcp-adapters. Use whenever a project needs a Grok 4 / GPT-class agent that talks to a live…