Agent
Performs Sunzian terrain-force-tempo analysis on any artifact. Maps environmental landscape, assesses capability-in-context, evaluates adaptation speed vs environmental change rate, and identifies strategic vulnerabilities. Decision - POSITIONED/EXPOSED.
Agent
Validates test quality after code passes the validator. Ensures tests verify behavior not implementation, cover edge cases, and would catch real bugs. Blocks progression if tests provide false confidence.
Agent
Analyzes whether agent thresholds produce the right signal-to-noise ratio. Consumes score distributions, false-positive rates, and convergence patterns to identify thresholds that are too lenient, too strict, or well-calibrated. Uses RAH statistical tools for sensitivity sweeps and robustness checks. Produces a…
Agent
Validates TypeScript type safety beyond compilation. Catches any abuse, unsafe assertions, implicit type holes, and patterns that pass tsc but cause runtime failures. Use AFTER code-validator for TypeScript projects. Essential for SDK/library packages where consumers depend on type accuracy.
Agent
Traces causal chains an artifact will unleash that weren't part of the design intent. Examines first-order effects, system interactions, and feedback loops. Actor-agnostic — consequences arise from any actor type through any mechanism. The broadest forecaster lens. Decision - ANTICIPATED/UNANTICIPATED.
Agent
Performs comprehensive knowledge-action unity analysis on any artifact. Maps the complete landscape of declared knowledge vs. enacted behavior. Produces a knowledge inventory, behavioral knowledge map, knowledge-action gap catalog with mechanism hypotheses (aspirational declaration, knowledge erosion, structural…
Agent
Performs Jamesian cash-value analysis on any artifact. Inventories distinctions, probes each for concrete lived difference in the experience and action of affected populations, applies the Lived-Action Test for evidence of differential action, distributes verdicts across populations via the Pluralism Check, and issues…
Agent
Performs Wittgensteinian language-game analysis on any artifact. Examines how terms are actually used across contexts, detects cross-game confusion where the same word plays different games, identifies pseudo-problems generated by conceptual confusion, and audits family-resemblance concepts governed by over-sharp…
Agent
Synthesizes cross-cutting insights from multiple upstream agent outputs in any workflow. Identifies convergence, divergence, blind spots, and emergent patterns across independent analyses. Produces meta-insights absent from any individual output. Decision - INTEGRATED/FRAGMENTED.
Command
Generates a coherent alternative frame for the problem an artifact solves — one that shares none of the artifact's assumptions. Not a competing solution but a competing understanding of the problem. Decision FRAMEAWARE/FRAMELOCKED.
Command
Finds terms and phrases used with multiple meanings within the same artifact. Produces a ranked ambiguity inventory with competing interpretations and disambiguation guidance.
Command
Reads from the position of someone afraid of the artifact's failure modes. Surfaces concerns hiding beneath confident language — classified by anxiety register (tactical/structural/epistemic). Decision CONFIDENCEWARRANTED/FRAGILITYMASKED.
Instructions file CodexOpenCode
AGENTS.md instructions for aself101/mcp-secure-server, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and layer implementation notes.
Instructions file
Claude Code instructions for aself101/mcp-secure-server, covering claude.md, project overview, commands, run all tests and run specific test suites.