Check it, do not trust it

Every claim shows its working

A coverage percentage is worth nothing if you cannot see what it rests on. Pick any released pair below and read the actual claims: which control does the work, what it satisfies, why, which document each control was verified against, on what date, who judged it, and whether the claim survived a pass whose job was to refute it.

Then read what did not survive. The mappings that were proposed for the same pair and rejected are published here too, with the reason each one failed. A coverage figure that never threw anything away has not been tested, and you cannot tell the difference from the number alone.

This is read live from the same endpoints your agent would call, and reading it here is free. An agent calls them with an API key or pays per call. Change the pair and watch it change. Nothing on this page is a picture of an audit trail.

598 released pairs to choose from.

57.4%
of the target already satisfied
35
of 61 controls evidenced
189
candidate mappings examined
102
removed in review, listed below

Level of review

machine_verified

Mappings were judged by Claude Code rather than read line by line by a practitioner. Every claim shows its reasoning so you can check it. Ask and a practitioner will review this pair.

25 claims, in full

GET /api/agent/crosswalk/provenance?source=ISO 27001:2022&target=SOC 2

Each claim names the source control whose evidence does the work, the target control it satisfies, and the reasoning. `argued_against_and_upheld` means a second pass tried to refute the claim and failed; claims that failed it are kept in the graph as refuted rather than deleted. The document fields say what each control was verified against and when.

8.6SOC2-A1.1high confidenceargued against and upheld
Evidence you already hold

Capacity management

What it satisfies

A1.1 Managing processing capacity

Reasoning

Capacity management is how availability commitments are kept.

Judged by Claude Code, Claude Max plan on 2026-08-19
8.13SOC2-A1.2high confidenceargued against and upheld
Evidence you already hold

Information backup

What it satisfies

A1.2 Environmental protection, backup and recovery infrastructure

Reasoning

Information backup is the data backup limb of this criterion.

Judged by Claude Code, Claude Max plan on 2026-08-19
7.11SOC2-A1.2high confidenceargued against and upheld
Evidence you already hold

Supporting utilities

What it satisfies

A1.2 Environmental protection, backup and recovery infrastructure

Reasoning

Supporting utilities cover the power and cooling this criterion names.

Judged by Claude Code, Claude Max plan on 2026-08-19
7.5SOC2-A1.2high confidenceargued against and upheld
Evidence you already hold

Protecting against physical and environmental threats

What it satisfies

A1.2 Environmental protection, backup and recovery infrastructure

Reasoning

Protecting against physical and environmental threats is the environmental limb.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.30SOC2-A1.3high confidenceargued against and upheld
Evidence you already hold

ICT readiness for business continuity

What it satisfies

A1.3 Testing recovery plan procedures

Reasoning

ICT readiness for continuity requires recovery plans be exercised.

Judged by Claude Code, Claude Max plan on 2026-08-19
8.11SOC2-C1.1high confidenceargued against and upheld
Evidence you already hold

Data masking

What it satisfies

C1.1 Identifying and maintaining confidential information

Reasoning

Data masking protects confidential information during processing and display.

Judged by Claude Code, Claude Max plan on 2026-08-19
6.6SOC2-C1.1high confidenceargued against and upheld
Evidence you already hold

Confidentiality or non-disclosure agreements

What it satisfies

C1.1 Identifying and maintaining confidential information

Reasoning

Confidentiality agreements bind those who handle confidential information.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.13SOC2-C1.1high confidenceargued against and upheld
Evidence you already hold

Labelling of information

What it satisfies

C1.1 Identifying and maintaining confidential information

Reasoning

Labelling makes classification actionable during handling.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.12SOC2-C1.1high confidenceargued against and upheld
Evidence you already hold

Classification of information

What it satisfies

C1.1 Identifying and maintaining confidential information

Reasoning

Classification of information is how confidential information gets identified.

Judged by Claude Code, Claude Max plan on 2026-08-19
8.10SOC2-C1.2high confidenceargued against and upheld
Evidence you already hold

Information deletion

What it satisfies

C1.2 Disposing of confidential information

Reasoning

Information deletion is the disposal requirement of this criterion.

Judged by Claude Code, Claude Max plan on 2026-08-19
7.14SOC2-C1.2high confidenceargued against and upheld
Evidence you already hold

Secure disposal or re-use of equipment

What it satisfies

C1.2 Disposing of confidential information

Reasoning

Secure disposal of equipment covers media holding confidential information.

Judged by Claude Code, Claude Max plan on 2026-08-19
6.4SOC2-CC1.1high confidenceargued against and upheld
Evidence you already hold

Disciplinary process

What it satisfies

CC1.1 Commitment to integrity and ethical values (COSO principle 1)

Reasoning

A disciplinary process is how deviations from expected conduct are addressed.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.3SOC2-CC1.3high confidenceargued against and upheld
Evidence you already hold

Segregation of duties

What it satisfies

CC1.3 Structures, reporting lines, authorities and responsibilities (COSO principle 3)

Reasoning

Segregation of duties defines reporting lines that limit conflicting authority.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.2SOC2-CC1.3high confidenceargued against and upheld
Evidence you already hold

Information security roles and responsibilities

What it satisfies

CC1.3 Structures, reporting lines, authorities and responsibilities (COSO principle 3)

Reasoning

Defined security roles and responsibilities establish structures and authorities.

Judged by Claude Code, Claude Max plan on 2026-08-19
6.3SOC2-CC1.4high confidenceargued against and upheld
Evidence you already hold

Information security awareness, education and training

What it satisfies

CC1.4 Attracting, developing and retaining competent people (COSO principle 4)

Reasoning

Awareness, education and training evidence developing and retaining competence.

Judged by Claude Code, Claude Max plan on 2026-08-19
6.1SOC2-CC1.4high confidenceargued against and upheld
Evidence you already hold

Screening

What it satisfies

CC1.4 Attracting, developing and retaining competent people (COSO principle 4)

Reasoning

Screening evidences the entity attracts individuals suited to their responsibilities.

Judged by Claude Code, Claude Max plan on 2026-08-19
6.4SOC2-CC1.5high confidenceargued against and upheld
Evidence you already hold

Disciplinary process

What it satisfies

CC1.5 Accountability for internal control responsibilities (COSO principle 5)

Reasoning

A disciplinary process holds individuals accountable for control responsibilities.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.4SOC2-CC1.5high confidenceargued against and upheld
Evidence you already hold

Management responsibilities

What it satisfies

CC1.5 Accountability for internal control responsibilities (COSO principle 5)

Reasoning

Management responsibilities require staff to apply security per defined rules.

Judged by Claude Code, Claude Max plan on 2026-08-19
6.3SOC2-CC2.2high confidenceargued against and upheld
Evidence you already hold

Information security awareness, education and training

What it satisfies

CC2.2 Internal communication of objectives and control responsibilities (COSO principle 14)

Reasoning

Awareness and training is how objectives and responsibilities reach personnel.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.1SOC2-CC2.2high confidenceargued against and upheld
Evidence you already hold

Policies for information security

What it satisfies

CC2.2 Internal communication of objectives and control responsibilities (COSO principle 14)

Reasoning

Communicated security policies carry objectives and responsibilities internally.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.2SOC2-CC2.2high confidenceargued against and upheld
Evidence you already hold

Information security roles and responsibilities

What it satisfies

CC2.2 Internal communication of objectives and control responsibilities (COSO principle 14)

Reasoning

Assigned roles tell individuals what they are internally responsible for.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.20SOC2-CC2.3high confidenceargued against and upheld
Evidence you already hold

Addressing information security within supplier agreements

What it satisfies

CC2.3 Communication with external parties about internal control (COSO principle 15)

Reasoning

Supplier agreements communicate security expectations to external parties.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.6SOC2-CC2.3high confidenceargued against and upheld
Evidence you already hold

Contact with special interest groups

What it satisfies

CC2.3 Communication with external parties about internal control (COSO principle 15)

Reasoning

Contact with special interest groups is external communication on security matters.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.5SOC2-CC2.3high confidenceargued against and upheld
Evidence you already hold

Contact with authorities

What it satisfies

CC2.3 Communication with external parties about internal control (COSO principle 15)

Reasoning

Contact with authorities is defined external communication about security matters.

Judged by Claude Code, Claude Max plan on 2026-08-19
5.36SOC2-CC4.1high confidenceargued against and upheld
Evidence you already hold

Compliance with policies, rules and standards for information security

What it satisfies

CC4.1 Ongoing and separate evaluations of control (COSO principle 16)

Reasoning

Compliance reviews against policies and standards are the ongoing evaluations.

Judged by Claude Code, Claude Max plan on 2026-08-19

102 mappings were refuted and are not in that number

GET /api/agent/crosswalk/refuted?source=ISO 27001:2022&target=SOC 2

These were proposed for this same pair and then thrown out. They are not deleted from the graph, they are kept marked as refuted, which is what stops a failed claim quietly coming back later. The coverage figure above is what is left after they were removed.

The pair summary counts every mapping removed in review. This section lists the ones marked refuted, which is the set that carries a record per mapping. Showing the first 25 of 102.

Read the sample as the best documented end of the list, not a fair slice of it. The endpoint returns the rejections that carry a stated reason first, and on several pairs most of the refutations behind them carry none. Raise the limit on the endpoint and you will see those too.

5.29SOC2-A1.2refuted, not countedclaimed medium confidence
Evidence it was proposed from

Information security during disruption

What it was claimed to satisfy

A1.2 Environmental protection, backup and recovery infrastructure

Why it failed

Cover backup, redundancy, continuity, and physical protections

5.30SOC2-A1.2refuted, not countedclaimed medium confidence
Evidence it was proposed from

ICT readiness for business continuity

What it was claimed to satisfy

A1.2 Environmental protection, backup and recovery infrastructure

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

8.14SOC2-A1.2refuted, not countedclaimed medium confidence
Evidence it was proposed from

Redundancy of information processing facilities

What it was claimed to satisfy

A1.2 Environmental protection, backup and recovery infrastructure

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

5.29SOC2-A1.3refuted, not countedclaimed high confidence
Evidence it was proposed from

Information security during disruption

What it was claimed to satisfy

A1.3 Testing recovery plan procedures

Why it failed

Held back 2026-08-19. Generated by the retired metered path, refuted at 56 percent wherever re-judged. Pair pulled from sale pending a rebuild on the Max plan.

8.13SOC2-A1.3refuted, not countedclaimed medium confidence
Evidence it was proposed from

Information backup

What it was claimed to satisfy

A1.3 Testing recovery plan procedures

Why it failed

5.30 and 5.29 cover testing and continuity during disruption

8.13SOC2-C1.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Information backup

What it was claimed to satisfy

C1.1 Identifying and maintaining confidential information

Why it failed

Classification, labeling, and transfer cover identification and protection

8.12SOC2-C1.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Data leakage prevention

What it was claimed to satisfy

C1.1 Identifying and maintaining confidential information

Why it failed

Classification, labeling, and transfer cover identification and protection

5.14SOC2-C1.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Information transfer

What it was claimed to satisfy

C1.1 Identifying and maintaining confidential information

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

6.6SOC2-CC1.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Confidentiality or non-disclosure agreements

What it was claimed to satisfy

CC1.1 Commitment to integrity and ethical values (COSO principle 1)

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

5.35SOC2-CC1.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Independent review of information security

What it was claimed to satisfy

CC1.1 Commitment to integrity and ethical values (COSO principle 1)

Why it failed

Policies, training, and NDAs show ethical commitment

5.2SOC2-CC1.1refuted, not countedclaimed high confidence
Evidence it was proposed from

Information security roles and responsibilities

What it was claimed to satisfy

CC1.1 Commitment to integrity and ethical values (COSO principle 1)

Why it failed

Policies, training, and NDAs show ethical commitment

5.36SOC2-CC1.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Compliance with policies, rules and standards for information security

What it was claimed to satisfy

CC1.1 Commitment to integrity and ethical values (COSO principle 1)

Why it failed

Policies, training, and NDAs show ethical commitment

6.3SOC2-CC1.1refuted, not countedclaimed low confidence
Evidence it was proposed from

Information security awareness, education and training

What it was claimed to satisfy

CC1.1 Commitment to integrity and ethical values (COSO principle 1)

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

5.35SOC2-CC1.2refuted, not countedclaimed high confidence
Evidence it was proposed from

Independent review of information security

What it was claimed to satisfy

CC1.2 Board independence and oversight of internal control (COSO principle 2)

Why it failed

Independent review not equivalent to board oversight

5.4SOC2-CC1.3refuted, not countedclaimed low confidence
Evidence it was proposed from

Management responsibilities

What it was claimed to satisfy

CC1.3 Structures, reporting lines, authorities and responsibilities (COSO principle 3)

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

5.2SOC2-CC1.5refuted, not countedclaimed high confidence
Evidence it was proposed from

Information security roles and responsibilities

What it was claimed to satisfy

CC1.5 Accountability for internal control responsibilities (COSO principle 5)

Why it failed

Held back 2026-08-19. Generated by the retired metered path, refuted at 56 percent wherever re-judged. Pair pulled from sale pending a rebuild on the Max plan.

6.3SOC2-CC1.5refuted, not countedclaimed medium confidence
Evidence it was proposed from

Information security awareness, education and training

What it was claimed to satisfy

CC1.5 Accountability for internal control responsibilities (COSO principle 5)

Why it failed

Roles, compliance, and discipline enforce accountability

5.36SOC2-CC1.5refuted, not countedclaimed medium confidence
Evidence it was proposed from

Compliance with policies, rules and standards for information security

What it was claimed to satisfy

CC1.5 Accountability for internal control responsibilities (COSO principle 5)

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

5.22SOC2-CC2.3refuted, not countedclaimed medium confidence
Evidence it was proposed from

Monitoring, review and change management of supplier services

What it was claimed to satisfy

CC2.3 Communication with external parties about internal control (COSO principle 15)

Why it failed

None address external communication about control functionality

6.4SOC2-CC3.3refuted, not countedclaimed medium confidence
Evidence it was proposed from

Disciplinary process

What it was claimed to satisfy

CC3.3 Considering fraud risk (COSO principle 8)

Why it failed

Neither addresses fraud risk assessment directly

5.24SOC2-CC3.3refuted, not countedclaimed medium confidence
Evidence it was proposed from

Information security incident management planning and preparation 

What it was claimed to satisfy

CC3.3 Considering fraud risk (COSO principle 8)

Why it failed

Neither addresses fraud risk assessment directly

8.16SOC2-CC4.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Monitoring activities

What it was claimed to satisfy

CC4.1 Ongoing and separate evaluations of control (COSO principle 16)

Why it failed

retired metered-API path. 56 percent of its output was refuted wherever it was re-judged, so its residue is not trusted evidence. Superseded by judgement on the Claude Max plan 2026-08-19.

8.32SOC2-CC4.1refuted, not countedclaimed medium confidence
Evidence it was proposed from

Change management

What it was claimed to satisfy

CC4.1 Ongoing and separate evaluations of control (COSO principle 16)

Why it failed

Ongoing monitoring and independent review directly support control evaluations

8.16SOC2-CC4.2refuted, not countedclaimed medium confidence
Evidence it was proposed from

Monitoring activities

What it was claimed to satisfy

CC4.2 Evaluating and communicating control deficiencies (COSO principle 17)

Why it failed

None ensure timely communication to senior management or board

5.26SOC2-CC4.2refuted, not countedclaimed high confidence
Evidence it was proposed from

Response to information security incidents

What it was claimed to satisfy

CC4.2 Evaluating and communicating control deficiencies (COSO principle 17)

Why it failed

None ensure timely communication to senior management or board

This is a sample, and the report is not

The endpoint returns up to twenty five claims so you can judge the work before paying for it. The paid report is every evidenced control and every gap, not a sample, as a report you can hand to an assessor.

What this is not

A mapping between two frameworks is a judgement, not text printed in either standard. Publishing the reasoning does not make a claim correct, it makes it checkable, and your assessor may still take a different view on individual controls. That is why the reasoning is here rather than a bare number.

Claims that failed the refutation pass are not deleted from the graph, they are kept marked as refuted, which is what stops a failed claim from quietly coming back. If you think one of the claims above is wrong, say so at support@theartofservice.com and it gets re-judged.