More screenshots
Works with the AI Agents you already use
ISO 27001 and SOC 2 Evidence Collection
Collect ISO 27001 / SOC 2 audit evidence like it matters: gap analysis first, API-first exports over screenshots, freshness windows, self-sorting filenames, and a control index.
$5
ISO 27001 and SOC 2 Evidence Collection
Example session with this skill installed
We're a 25-person startup. Our SOC 2 Type II observation window starts in two months and we have collected exactly nothing. Where do we start, and what does the first pass look like?
- Read your context and instructions
- Compiled the iso 27001
FIRST-PASS PLAN — produced by applying the skill.
Week 0, gap analysis before collecting anything: your criteria list (say 61 applicable controls across the five trust services categories) gets three marks each — current / stale / missing. With nothing collected, expect most to read "missing," but the walk still pays: it separates "we do it and can prove it" from "we do it and can't prove it" from "we don't do it." Only the third group is a control problem; the second is a collection problem, and it's most of the list at your size.
Collection order follows the protocol: critical-and-missing first. Day one, send the vendor requests — your cloud host, payments, and monitoring vendors' SOC 2 reports take weeks and cannot be rushed later. Then platform sessions, one per system, API-first: identity provider (user list with MFA status, admin assignments, leavers in the period — level 1 JSON exports, not screenshots); code hosting (org MFA enforcement, branch protection on production repos, merged-PR records for change management); cloud (IAM policies, audit-log configuration, backup configuration, encryption at rest).
Filenames from the first file onward: CC6.1_idp-user-mfa_2026-10-23.json — control first, date last, so the folder sorts itself. Index rows as you go: control | file | hierarchy level | collected date | status.
The honest warning the skill insists on: your observation window starts in two months, and Type II judges operation over the period. Evidence you start collecting now will be strong; anything that should have been happening quarterly (access reviews) needs to start this week so the window contains at least one real instance — and the gap before it gets documented, not backfilled. Fabrication ends audits; a dated gap note with a running cadence reads as a control maturing.
Readiness target before the window opens: rubric 8+/10 with no zero on coverage or freshness.
Connects securely to your tools. The creator never sees your data.
About this skill
Audits are won or lost before the auditor arrives, in the quality of the evidence package — and most packages fail in predictable ways: undated screenshots, console output pasted into documents, a current access list offered as proof that quarterly reviews happened. This skill is the collection discipline that avoids all three, for ISO 27001:2022 and SOC 2 alike. It starts with an evidence hierarchy — API/CLI exports with timestamps at the top, then system-generated reports, configuration exports, screenshots with a visible system clock, and signed attestations as a last resort that needs corroboration — and a rule: before reaching for a weak level, check whether a stronger source exists untried. Freshness gets a table, because evidence expires: vulnerability scans at 30 days, access reviews and config exports at 90, training records and penetration tests at 12 months, incident records continuous across the whole audit period. The protocol runs gap analysis first (every control marked current, stale, or missing, collected in blocking order), then collects by platform rather than by control — identity provider, code hosting, cloud, endpoints, vendors — one session per system pulling everything it owes, with vendor SOC 2 reports requested on day one because vendors are slow. Filenames follow a convention that sorts by control and then by date, so the package assembles itself. The index maps every control to its files, hierarchy level, dates, and status, and it's honest: gaps marked as gaps, because a self-declared gap with a corrective plan reads as an observation while a discovered one reads as a finding. The skill includes a gap/index template, a per-platform session checklist, and a readiness rubric across coverage, freshness, strength, traceability, and integrity — hand over at 8 of 10, never with a zero on coverage or freshness. A worked example walks a 40-person SaaS company six weeks out from its first Stage 1: 13 missing controls clustered in access reviews and vendor reports, two unrecoverable quarters disclosed rather than fabricated, and a 9/10 package that drew one observation instead of a nonconformity.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.
- It collects and validates proof; it does not implement controls, remediate gaps it finds, or replace your auditor's or legal counsel's judgment on scope.
- Freshness windows and period conventions reflect common auditor practice for ISO 27001:2022 and SOC 2; your certification body's specific requirements govern where they differ.
- Evidence for past periods cannot be created retroactively — where the skill finds an unrecoverable gap, its answer is an honest gap note and a cadence, not a workaround.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 24 days ago
Permissions
No special permissions declared or detected