Deep Trajectory Analysis

Reconstruct and visually analyze paired agent, game, or policy trajectories to determine whether changed actions produced their intended effects. Use for move-history audits, replay/trace analysis, policy regressions, behavior calibration, causal first-divergence studies, reply-survival analysis, agent personality validation, promotion decisions, and reports that must connect aggregate outcomes to exact state-action sequences.

201 stars 12 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add swyxio/skills --skill deep-trajectory-analysis
Manual install
git clone --depth 1 https://github.com/swyxio/skills.git /tmp/skills
mkdir -p ~/.claude/skills
cp -r /tmp/skills/deep-trajectory-analysis ~/.claude/skills/deep-trajectory-analysis

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Deep Trajectory Analysis

Determine not merely whether a candidate scored better, but what it changed, whether the named
mechanism fired, how the environment replied, and whether the benefit survived to the final
outcome.

Read [references/evidence-contract.md](references/evidence-contract.md) before defining metrics or
building the report. Run [scripts/validate-report.mjs](scripts/validate-report.mjs) on a
machine-readable result that follows that contract.

## Establish the causal question

Write one sentence for each intended effect before inspecting favorable examples:

> When opportunity O is publicly observable, policy P should change action A, immediately alter
> mechanism M, survive response R, and improve horizon outcome H without violating safety S.

Operationalize every noun. Prefer board-, state-, or trace-derived quantities over labels such as
“aggressive,” “safe,” “precise,” or “wild.”

Separate:

- **strength**: final win, reward, score margin, rank, or task success;
- **identity**: the behavior the policy is meant to exhibit;
- **safety**: compute, fairness, legality, information access, termination, and worst-case floors.

Do not let one metric stand in for all three.

## Preflight integrity

Stop promotion analysis until these checks pass:

1. Freeze engine/model, rules, maps/environments, search budget, opponent/control policies, seeds,
   seats/roles, starter/order, horizon, and terminal scoring.
Read the full skill on GitHub

Details

Skill name
deep-trajectory-analysis
Repository
swyxio/skills
Path
deep-trajectory-analysis/SKILL.md
License
MIT
Repo updated
16 November 2026

FAQ

What does the Deep Trajectory Analysis skill do?

Reconstruct and visually analyze paired agent, game, or policy trajectories to determine whether changed actions produced their intended effects. Use for move-history audits, replay/trace analysis, policy regressions, behavior calibration, causal first-divergence studies, reply-survival analysis, agent personality validation, promotion decisions, and reports that must connect aggregate outcomes to exact state-action sequences.

How do I install the Deep Trajectory Analysis skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add swyxio/skills --skill deep-trajectory-analysis. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Deep Trajectory Analysis popular?

It lives in swyxio/skills, which has 201 GitHub stars and 12 forks.

Related skills

Accelevents API

201

swyxio/skills

Use when reading or updating AI Engineer Europe speaker records through the Accelevents REST API, especially for full-record PUT updates, auth header quirks, and preserving existing speaker fields.

Software Development12

Accelevents Speaker Sync

201

swyxio/skills

Use when website speaker, session, schedule, room, track, or headshot changes must be synchronized back to Accelevents for AI Engineer Europe 2026.

AI, LLMs & Agents12

Agents SDK

201

swyxio/skills

Build or debug stateful applications that specifically use Cloudflare's `agents` package, including Agent classes, routing, callable methods, synchronized state, scheduling, AIChatAgent, durable execution, or Agents SDK client hooks. Do not trigger for generic AI-agent design, the OpenAI Agents SDK, or ordinary Durable Object work that does not use the Cloudflare Agents SDK.

AI, LLMs & Agents12

Design Trends 2026

150

guia-matthieu/clawfu-skills

Apply 2026's top graphic design trends to any creative brief. Based on Kittl × Savee's 2026 Design Trends Report (10 trends + 2 honorable mentions), backed by Adobe, Figma, and Pinterest data. Use when: **Designing a brand identity** — pick the right aesthetic for your audience; **Creating social media assets** — use trending visual languages that perform; **Briefing a designer or AI image tool** — give precise style direction with vocabulary and references; **Refreshing a visual identity** — know what's rising vs saturating; **Building mood boards** — combine trends intentionally with data-ba

AI, LLMs & AgentsIn 1 list32

Data Chatbots

201

swyxio/skills

Design and implement AI copilots that analyze structured data and propose mutations for human approval (draft → apply), not direct writes. Covers prompting, multistep compound requests (remove + add in one turn), FIFO client queuing for long agent loops, validation, session memory, optimistic versioning UX, copilot panel/surface ergonomics (floating-dockable panel, summon-hotkey size state machine), LLM tracing/observability (OpenInference + OTLP session.id grouping, orphan-span fixes, human thumbs up/down feedback annotations), and test matrices. Use when building scheduling/admin chatbots, p

AI, LLMs & Agents12

Deterministic Metric Design

228

pproenca/dot-skills

Inventing deterministic metrics — turning a fuzzy property like 'maintainability', 'risk', or 'how reducible this code is' into a deterministic, computable number an agent can trust and optimize. Covers the path from construct to adoption — operationalizing the construct, confronting computability limits (Kolmogorov, Rice) with sound proxies, picking the right measurement scale, proving properties (monotonicity, invariance, the Weyuker/Briand axioms), guaranteeing determinism, establishing construct validity (not just LOC in disguise), and hardening against Goodhart-style gaming when an agent

AI, LLMs & Agents19

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.