“I know Python” is a claim.

A scored Story Quest attempt under stated constraints is a different signal. Most career surfaces blur them into one confident line on a profile. Reviewers then guess how much trust to assign—or fall back to degrees.

An attempt is behaviour. A skill claim is a capability signal someone else can weight at a stated tier.

This guide owns the trust ladder: how practice upgrades a claim from self-declared to demonstrated. It does not rehash what belongs in a career evidence record, why portfolios lose context, or how the Shared Footprint loop closes—point once, then stay on tiers.

Assertion is not substantiation

Education Design Lab defines skills validation as the process by which an assertion (“I assert that I have a skill!”) is substantiated—typically by a qualified third party, against a shared rubric, with level and context, via multiple methods (automated assessments, social endorsements, other approaches).

Their Skills Validation Guidebook (2024) states the hiring failure mode cleanly: without validation, employers rely on proxies such as degrees, while candidates’ skills are “often simply asserted.” Skills must be validated to establish trust and offer demonstrations of what a candidate knows and can do.

Tiers exist so a skim does not treat every skill chip as equal.

Four verification tiers

ProfilEd language aligned to standards vocabulary—not a copy of any wallet UI.

Tiers can stack. They are not mutually exclusive. A claim can be self-declared and attested and demonstrated.

What each tier is not

• Verified ≠ “we like this person.” In Open Badges 3.0, verification means the credential conforms, the proof method is satisfied, and status checks succeed. It is not automatic validation of job fitness.
• Attested ≠ assessment. An endorsement concurs with a claim; it does not replace scored performance.
• Demonstrated can exist without a formal badge. A badge without performance evidence stays closer to declared or issuer-verified claim than to demonstrated capability.
• Self-declared is legitimate when labeled. The damage starts when it wears a verified costume.

Open Badges 3.0 already separates Skill Claim, Evidence (supporting information such as a URL to an artefact), Verification, Self-assertion, and Endorsement. Verifiers are told to trust the issuer—including the risk that a subject can self-issue. HR Open Standards’ Trusted Career Profile deliberately mixes LERs/VCs (Open Badges, CLRs) with self-attested skills, history, and narratives in one portable package. Mixed trust density is expected. Unlabeled mixes fake trust.

How practice upgrades a claim

Learner action before AI. Practice writes the attempt. AI may probe. Credentials and endorsements sit on top of the trail—not instead of it.

U.S. OPM frames work samples and simulations as tasks that “mirror the tasks employees perform on the job,” with high content and criterion-related validity. Classic meta-analysis treated work samples among the strongest predictor types (Schmidt & Hunter, 1998). Sackett and colleagues (2022) revised operational validity for work samples downward (often summarised near .33) after correcting range-restriction overcorrection—still top-tier among job-specific predictors. Prefer type-level language if you skip the pair of citations.

Vendor ladders that grade badges from “Experience” (unmeasured) through “Validation” and “Certification” (Credly guidance) rhyme with measurement intensity. Rename into the four tiers above; do not copy product UI.

One skill, climbing the ladder

Take a single Graph goal: stakeholder negotiation under time pressure.

1. Self-declared — learner tags the skill on the Curriculum Graph.
2. Demonstrated — completes a Story Quest scenario; attempt, rubric score, and decision trail land in Shared Footprint.
3. Attested — a mentor endorses that footprint slice.
4. Verified — an issuer (or institutional badge) attaches an Open Badge / CLR-aligned credential with criteria and evidence URL pointing at the same trail.

A reviewer can now weight each chip honestly: declaration, vouch, cryptographic credential, and observed performance are no longer one blurry “proficient.”

Reviewer skim contract

Label the tier on every skill signal. Prefer demonstrated and attested over unlabeled self-report when time is scarce. Use verified credentials for integrity of issuance—not as a substitute for watching the work.

Assessment-only platforms that stop at a score, and proof wallets that only chain artifacts after the fact, both miss the upgrade path: learner-first attempts that feed a named tier inside a Shared Footprint.

Start labeling trust

Turn attempts into labeled skill claims on ProfilEd →

Sources (research brief only)

1. Education Design Lab — skills validation glossary (2023)
2. Education Design Lab — Skills Validation Guidebook (2024)
3. 1EdTech Open Badges Specification v3.0
4. HR Open Standards — Trusted Career Profile
5. U.S. OPM — Work Samples and Simulations
6. Schmidt & Hunter (1998); Sackett et al. (2022)
7. Credly issuer metadata guidance (vendor — complementary)

Ops note

• Gap: CMS route tutorials vs insights still open for skills-evidence flagship.
• Improvement: Lock skills-evidence category + four-tier table as reusable Directus block.

From attempt to skill claim (verification tiers)

“I know Python” is a claim.

A scored Story Quest attempt under stated constraints is a different signal. Most career surfaces blur them into one confident line on a profile. Reviewers then guess how much trust to assign—or fall back to degrees.

An attempt is behaviour. A skill claim is a capability signal someone else can weight at a stated tier.

This guide owns the trust ladder: how practice upgrades a claim from self-declared to demonstrated. It does not rehash what belongs in a career evidence record, why portfolios lose context, or how the Shared Footprint loop closes—point once, then stay on tiers.

Assertion is not substantiation

Education Design Lab defines skills validation as the process by which an assertion (“I assert that I have a skill!”) is substantiated—typically by a qualified third party, against a shared rubric, with level and context, via multiple methods (automated assessments, social endorsements, other approaches).

Their Skills Validation Guidebook (2024) states the hiring failure mode cleanly: without validation, employers rely on proxies such as degrees, while candidates’ skills are “often simply asserted.” Skills must be validated to establish trust and offer demonstrations of what a candidate knows and can do.

Tiers exist so a skim does not treat every skill chip as equal.

Four verification tiers

ProfilEd language aligned to standards vocabulary—not a copy of any wallet UI.

Tiers can stack. They are not mutually exclusive. A claim can be self-declared and attested and demonstrated.

What each tier is not

• Verified ≠ “we like this person.” In Open Badges 3.0, verification means the credential conforms, the proof method is satisfied, and status checks succeed. It is not automatic validation of job fitness.
• Attested ≠ assessment. An endorsement concurs with a claim; it does not replace scored performance.
• Demonstrated can exist without a formal badge. A badge without performance evidence stays closer to declared or issuer-verified claim than to demonstrated capability.
• Self-declared is legitimate when labeled. The damage starts when it wears a verified costume.

Open Badges 3.0 already separates Skill Claim, Evidence (supporting information such as a URL to an artefact), Verification, Self-assertion, and Endorsement. Verifiers are told to trust the issuer—including the risk that a subject can self-issue. HR Open Standards’ Trusted Career Profile deliberately mixes LERs/VCs (Open Badges, CLRs) with self-attested skills, history, and narratives in one portable package. Mixed trust density is expected. Unlabeled mixes fake trust.

How practice upgrades a claim

Learner action before AI. Practice writes the attempt. AI may probe. Credentials and endorsements sit on top of the trail—not instead of it.

U.S. OPM frames work samples and simulations as tasks that “mirror the tasks employees perform on the job,” with high content and criterion-related validity. Classic meta-analysis treated work samples among the strongest predictor types (Schmidt & Hunter, 1998). Sackett and colleagues (2022) revised operational validity for work samples downward (often summarised near .33) after correcting range-restriction overcorrection—still top-tier among job-specific predictors. Prefer type-level language if you skip the pair of citations.

Vendor ladders that grade badges from “Experience” (unmeasured) through “Validation” and “Certification” (Credly guidance) rhyme with measurement intensity. Rename into the four tiers above; do not copy product UI.

One skill, climbing the ladder

Take a single Graph goal: stakeholder negotiation under time pressure.

1. Self-declared — learner tags the skill on the Curriculum Graph.
2. Demonstrated — completes a Story Quest scenario; attempt, rubric score, and decision trail land in Shared Footprint.
3. Attested — a mentor endorses that footprint slice.
4. Verified — an issuer (or institutional badge) attaches an Open Badge / CLR-aligned credential with criteria and evidence URL pointing at the same trail.

A reviewer can now weight each chip honestly: declaration, vouch, cryptographic credential, and observed performance are no longer one blurry “proficient.”

Reviewer skim contract

Label the tier on every skill signal. Prefer demonstrated and attested over unlabeled self-report when time is scarce. Use verified credentials for integrity of issuance—not as a substitute for watching the work.

Assessment-only platforms that stop at a score, and proof wallets that only chain artifacts after the fact, both miss the upgrade path: learner-first attempts that feed a named tier inside a Shared Footprint.

Start labeling trust

Turn attempts into labeled skill claims on ProfilEd →

Sources (research brief only)

1. Education Design Lab — skills validation glossary (2023)
2. Education Design Lab — Skills Validation Guidebook (2024)
3. 1EdTech Open Badges Specification v3.0
4. HR Open Standards — Trusted Career Profile
5. U.S. OPM — Work Samples and Simulations
6. Schmidt & Hunter (1998); Sackett et al. (2022)
7. Credly issuer metadata guidance (vendor — complementary)

Ops note

• Gap: CMS route tutorials vs insights still open for skills-evidence flagship.
• Improvement: Lock skills-evidence category + four-tier table as reusable Directus block.