strategic-analyst

Produce a deep strategic brief on a topic — provenance check, multi-source fetch, evidence ledger, multi-lens analysis, calibrated synthesis, red-team review.

Verify a single claim or URL — chain of custody, two-source rule, primary fetch, archive, and Admiralty-graded ledger row with verdict.

Sweep the configured watchlist of named outlets, dedupe against prior runs, cluster by topic, and produce today's morning situation report.

Stress-test an existing assessment — Key Assumptions Check, alternative hypotheses, weakest evidence, pre-mortem, calibration adjustment.

Use when applying a single analytical lens (economic, political, historical, demographic, military, public-finance, or geographic) to one event. Designed to be invoked in parallel — one agent per lens that has been triaged as load-bearing. Returns a structured finding (claim with calibration, evidence weight, residual uncertainty, citations), not narrative. <example> Context: orchestrator has triaged that public-finance, economic, and political lenses are load-bearing for an Argentine capital-controls event. user: [orchestrator dispatches three lens-applier agents in parallel] assistant: lens-applier returns three structured findings; orchestrator merges them in synthesising-strategic-assessment </example> <example> Context: a sovereign-debt downgrade story. user: "apply the public-finance lens to the Bolivia downgrade" assistant: invoke lens-applier with lens=public-finance and the event description; receive a structured finding back </example>

Use to stress-test a drafted strategic assessment before it goes to a principal. Designed to be invoked as a separate agent so the review is structurally independent of the drafting reasoning — the red-teamer sees only the finished assessment plus the evidence ledger, not how the analyst got there. Returns a structured review with calibration adjustments. The /red-team slash command invokes this agent. <example> Context: an analyst has drafted briefs/turkey-cbrt-decision-2026-05-05.md with the corresponding evidence/turkey-cbrt-decision-2026-05-05.csv ledger. user: invokes /red-team assistant: dispatches red-teamer with the assessment path and ledger path; receives a structured review that adjusts the calibration band and adds indicators </example> <example> Context: a synthesis just completed in-thread; the user wants an independent check before sending. user: "red-team this" assistant: dispatches red-teamer with the assessment text; receives a fresh adversarial review uncontaminated by the drafting reasoning </example>

Use to ingest a large set of articles, feed items, or URLs and return a compressed, graded summary that the main thread can consume without burning context on raw text. Designed to keep article bodies out of the orchestrator's window — fetches and reads in its own context, returns narrative clusters plus evidence-ledger rows. Invoked by /strategic-brief and /daily-sitrep when N > ~10 sources are involved. <example> Context: orchestrator wants the last 24h of global coverage of "Turkey CBRT rate decision" without reading 80 articles in the main thread. user: "ingest the last 24h of news on this topic and tell me what's there" assistant: dispatches source-ingestor with topic + window; receives 3 narrative clusters + 8 ledger rows; never sees the article bodies </example> <example> Context: a /daily-sitrep run pulls 200 RSS items across 14 outlets. user: invokes /daily-sitrep assistant: dispatches source-ingestor with the watchlist + dedup state; gets back the day's narratives + new ledger rows; assembles the sitrep from those, not from the raw items </example>

Use when an event implicates population structure, fertility, migration, ageing, urbanisation, ethnic composition, or labour supply. Triggers include "ageing population", "youth bulge", "migration crisis", pension or health-system stress, ethnic or sectarian dynamics, demographic dividend or drag.

Use when interpreting currency moves, trade flows, inflation prints, growth data, sanctions, supply-chain shocks, commodity moves, or central-bank actions. Triggers include "what's the economic impact", "why did markets move", "is this inflationary", "who wins from this trade policy", or any claim of economic causation.

Use when an event implicates terrain, distance, chokepoints, resources, climate, or maritime or pipeline geography. Triggers include shipping-lane disruptions, pipeline politics, border or territorial questions, infrastructure projects, satellite-derived claims, or "why does X matter strategically".

Use when placing an event in historical context — precedent, analogies, cycles, path dependencies, or claims that something is "unprecedented". Triggers include "has this happened before", "is this like X", "what's the precedent", long-run trend reasoning, or structural-break vs continuation calls.

Use when an event involves armed forces, deterrence, alliances, procurement, escalation risk, or grey-zone activity. Triggers include troop movements, exercises, weapons announcements, missile tests, naval deployments, cyber operations, "is this escalation", or "what's the military balance".

1 hook across 1 event