SAVVYDFIR-MCP: Agent for Claude Code

.claude/agents/prefetch-analyst.md

prefetch-analyst is an agent for Claude Code from kismatkunwar89/SAVVYDFIR-MCP. It costs 72 tokens per session (3,379 once invoked), scanned A, original, MIT.

A forensic analysis add-on for Windows Prefetch files, which record information about applications run on a Windows computer. It looks for execution patterns and signs that the records may have been abused or removed.

In plain words
What is it for?
Use it to detect execution from multiple paths, identify SysWOW64 and other built-in-tool abuse, examine referenced files, investigate orphaned Prefetch records, and find possible Prefetch deletion.
Why use it?
It helps investigators determine how and when programs ran, including suspicious execution paths and attempts to hide activity. It also connects Prefetch evidence with related files and movements.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

This is kismatkunwar89/SAVVYDFIR-MCP's own configuration. It tells Claude Code how to work on SAVVYDFIR-MCP itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything SAVVYDFIR-MCP configures →

Reuse

Borrowing it

Nothing to install: this file belongs to kismatkunwar89/SAVVYDFIR-MCP. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/kismatkunwar89/SAVVYDFIR-MCP/master/.claude/agents/prefetch-analyst.md
Clone the repo
git clone --depth 1 https://github.com/kismatkunwar89/SAVVYDFIR-MCP

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for prefetch-analyst

README.md
[![agentmods](https://agentmods.dev/badge/agents/kismatkunwar89/savvydfir-mcp/prefetch-analyst/github.svg)](https://agentmods.dev/agents/kismatkunwar89/savvydfir-mcp/prefetch-analyst)
Your own site
<a href="https://agentmods.dev/agents/kismatkunwar89/savvydfir-mcp/prefetch-analyst"><img src="https://agentmods.dev/badge/agents/kismatkunwar89/savvydfir-mcp/prefetch-analyst/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for prefetch-analyst

Your own site · 80×15
<a href="https://agentmods.dev/agents/kismatkunwar89/savvydfir-mcp/prefetch-analyst"><img src="https://agentmods.dev/badge/agents/kismatkunwar89/savvydfir-mcp/prefetch-analyst.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 72 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,379 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00072 $0.03379
Opus 5 $0.00036 $0.01690
Sonnet 5 $0.00014 $0.00676
Haiku 4.5 $0.00007 $0.00338

Measured 10d ago against content hash 561b7087f071, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

prefetch-analyst 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 10d ago.

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.

.claude/agents/prefetch-analyst.md · 262 lines

How it starts

The opening of the file, as written. The whole thing — 262 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Windows Prefetch Forensic Analyst

How this file is used

This is a forensic-heuristic knowledge base, not a procedural playbook. The main investigator agent reads this file as reference context when analyzing the relevant artifact. Apply heuristics where they fit the case context - do not execute them as a fixed sequence.

For court-defensible findings: cite the specific tool execution and raw evidence that supports each claim. Use submit_finding with structured provenance (execution_id, evidence_excerpt, contradictions, corroborations).

The user-authored heuristics below were preserved verbatim during the 2026-05-23 Phase 3 overlay removal.

Forensic Ground Rules

  • NEVER load raw CSV rows into context - write targeted Pandas queries via run_analysis only
  • Schema discovery is mandatory first - pyscca column names vary by parser version
  • Every confirmed anomaly gets an immediate add_finding call before the next query
  • Call read_state first for case status and attack-window summary, then call get_findings when you need the full prior MFT/EVTX finding set

What Prefetch Tells You

A .pf file is created when Windows executes an application - proving a binary actually ran, not just existed. Key fields:

  • ExecutableName: the binary name that ran
  • RunCount: exactly how many times it executed (may be capped)
  • LastRunTimes: up to 8 most recent execution timestamps (subtract ~10 seconds - file is written after monitoring period)
  • FilesLoadedList / file_references: every file, DLL, and directory the application touched in its first 10 seconds
  • Creation timestamp (filesystem): first-ever execution; Modification timestamp: most recent execution
  • Hash in filename (e.g., CMD.EXE-1A2B3C4D.pf): calculated from the executable's full directory path (exception: svchost, rundll32, dllhost, mmc include command-line args in hash)

Critical Forensic Heuristics

Use your forensic training - extend beyond these indicators based on what the data reveals:

Read the full file on GitHub · 262 lines

Changes

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.

  1. 10d ago First seen · 262 lines · 72 tokens per session scan A 561b7087f071

Subscribe to this mod's changes

prefetch-analyst is an agent published in the GitHub repository kismatkunwar89/SAVVYDFIR-MCP (4 stars, last pushed 2mo ago), licensed MIT. It adds 72 tokens to every session and 3,379 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 2026-08-31.

Related

Other agents, from other repositories