Skill

analyzing-browser-forensics-with-hindsight

skill:mukul975/anthropic-cybersecurity-skills#analyzing-browser-forensics-with-hindsight

Parse Chromium-based browser databases with Hindsight to extract and correlate browsing history, downloads, cookies, cached content, autofill data, saved passwords, and extensions from Chrome, Edge, Brave, Opera, and Vivaldi into a unified timeline (XLSX, JSON, or SQLite output). Use during incident response, insider-threat investigations, or criminal cases when you need to reconstruct a user's web activity from a browser profile.

skills.sh SkillsMP

Verdict

Suspicious

Risky behavior the analysis could not rule out.

Analyzed Oct 27, 2026, 20:54 UTC at 54a7988

findings

medium

highest severity

threat classes

4,532

files scanned

135/135

checks satisfied

package advisories

Why this verdict

suspicious: finding skill:skills/analyzing-browser-forensics-with-hindsight:skills/analyzing-browser-forensics-with-hindsight/SKILL.md:47:SUPPLY

Security findings 1

Supply chain

mediumContextual risklow confidence

The skill's prerequisites instruct an unpinned install of the Hindsight dependency (`pip install pyhindsight`), leaving the code fetched at install time mutable; the package name matches the official distribution and the install is a genuine prerequisite, so this is an unmitigated supply-chain surface rather than evidence of abuse.

skills/analyzing-browser-forensics-with-hindsight/SKILL.md:47

Python 3.8+ with Hindsight installed (`pip install pyhindsight`)

vector:instruction

Counter-argument considered

The quote is exact and the install is unpinned, so the pinned-official-registry carve-out does not apply and no mitigation appears in any shown unit, leaving a genuine but minor supply-chain surface; the strongest counter-argument is that the instruction is overt and matches the declared 'with Hindsight' purpose, which is reflected in the failed undeclared gate but caps the finding at its own contextual_risk/low scope rather than making it wrong.

Refuted findings 1

Candidates the review stage examined and rejected, kept for transparency.

Credential access

criticalRefutedmedium confidence

The skill's shipped agent code and its reference docs access browser credential-adjacent stores — reading the Cookies database (host, session-cookie names, expiry, secure/httponly flags) and autofill user-entered values, with reference documentation instructing a SELECT that includes the plaintext cookie value column — emitting this data into a local console/JSON report, which exposes session identifiers and form-entered personal data on the analyzed host but matches the skill's declared forensic purpose and has no off-host sink in the shown evidence.

skills/analyzing-browser-forensics-with-hindsight/scripts/agent.py:143

SELECT name, value, count, date_created, date_last_used

Why it was refuted

Cookie and autofill access is the declared core capability: the skill's declared description explicitly lists 'cookies' and 'autofill data' among what it extracts, the shipped reference documentation presents these queries as the skill's open methodology, and the shown code uses read-only sqlite access with local-only returns and no off-host sink, so the declared-core-capability carve-out defeats the harm and undeclared gates and the finding is refuted.