Plugin Claude Code
FOSS-first investigative analysis toolkit. Turn FOIA releases and data dumps into findings you can stand behind: counted, cited to the exact page, swept for PII, and filed as linked notes, all locally in your own Claude Code session. Covers structured-data analysis, document ingest, redaction checks, and opt-in entity.
MCP server Claude CodeCodexCursor +2
MCP server for SQLite files. Supports Datasette-compatible metadata! Runs locally from the mcp-sqlite Python package.
Agent
Use this agent during the investigate verification gate to mechanically check every extracted claim's citation anchor by driving the deterministic resolver in scripts/citation.py (resolveanchor + iscleancitation). It reports a per-claim {anchorlevel, ok, reason} and flags uncited claims, degraded anchors…
Agent
Use this agent during the investigate verification gate to perform a blinded, adversarial semantic re-read of ONE extracted claim against the exact source span the extractor cited. It runs a presence check and an entailment check and returns an advisory supported/contradicted/indeterminate verdict for the human…
Skill Claude CodeCodex
This skill should be used when the user asks to "run the analysis recipe", "run the 13-point checklist", "run the per-source investigative pass", "analyze this FOIA/audit-log source for out-of-state / immigration / pretext / mega-users / co-travel", "roll up findings across multiple sources/agencies", or "test the…
Skill Claude CodeCodex
Use when a user receives a FOIA artifact (a released record, a document dump, an export) and wants to record its provenance and establish a chain of custody before any processing. This skill hashes the file on receipt, obtains an RFC 3161 trusted timestamp proving that hash existed at a time, writes an append-only…
Skill Claude CodeCodex
This skill should be used when the user asks to "analyze a FOIA dataset", "load and analyze an audit-log CSV/XLSX", "build a queryable dataset", "run the dataset analysis pipeline", "check a data export for truncation", or otherwise wants to turn a dirty FOIA / audit-log CSV or spreadsheet into clean, derived…
Skill Claude CodeCodex
Use when a journalist or investigator wants to check whether Magpie is healthy on this machine -- which analysis and document workflows are ready, what is missing, and what to ask their operator to fix. This is a strictly read-only health check; it installs nothing. Invoke it whenever a user wants to run a health…
Skill Claude CodeCodex
Cross-reference an investigation's resolved entities against the investigator's own corpus (FOSS default, zero external data) and, opt-in, sanctions/PEP watchlists, using a local yente + OpenSearch stack. Layer-2, operator-tier, Docker-gated -- run it after building the entity graph to screen the resolved people/orgs…
Skill Claude CodeCodex
This skill should be used when the user wants to extract named entities and the relations between them from a trustworthy ingested document, build an entity/relationship network or knowledge graph from a FOIA PDF, map people / organizations / agencies / vendors and their connections, or produce a reviewed FtM-shaped…
Skill Claude CodeCodex
Resolve entities across an investigation's documents into a Neo4j graph after a mandatory human review gate. Layer-2, operator-tier, Docker-gated -- run it to deduplicate people/orgs across the corpus, decide the "maybe" matches yourself, then write the resolved network to Neo4j for queryable analysis.
Skill Claude CodeCodex
This skill should be used when the user asks to "ingest a PDF / document", "extract text from a scanned PDF preserving page and bounding-box provenance", "convert a FOIA PDF to DoclingDocument keeping citations", "OCR a scanned document but decide native-text-vs-re-OCR first", "turn a PDF into provenance-preserving…
Skill Claude CodeCodex
This skill should be used when the user wants to investigate a trustworthy ingested document, extract findings or claims from it, turn a converted PDF into cited findings, fact-check or verify AI-extracted claims against the source, or publish anchored findings to the Librarian. It runs the verification gate that…
Skill Claude CodeCodex
This skill should be used when the user asks to "sweep PII exposure in a FOIA free-text field", "quantify how much PII a free-text reason/narrative column exposed", "run the authoritative PII tally", "count names / SSNs / A-numbers / DOBs in a reason column weighted by row count", "separate officials named for…
Skill Claude CodeCodex
This skill should be used when the user asks to "redact PII before publishing", "redact third-party names to initials in a findings note", "redact a note before it goes in the vault", "mask suspect / person-of-interest / minor names that landed in the reason fields", "keep the full un-redacted exhibit local / out of…
Skill Claude CodeCodex
This skill should be used when the user asks to "check a PDF for bad redactions", "find improper / failed redactions", "did this FOIA response leak text under the black boxes", "is there still-extractable text under the redaction boxes", "scan a PDF for un-redacted metadata / embedded files / form values", "check my…
Skill Claude CodeCodex
Use when an operator is setting up Magpie on a new machine for the first time and needs to install the Python dependencies, download the spaCy model, and verify the local toolchain. This skill runs the repo-managed bootstrap, instructs the operator for the system binaries pip cannot install, and re-checks the…