Self-assessment, not a certification. The published pages of uncovertechtalent.com are the output of one deployed assembly (a model, its instruction files, hooks, memory, a knowledge vault and a human operator). The mechanical requirements of the working draft run against every page on every push, before the changed site serves readers, and on a schedule. This page is rendered from the latest run record.
A mechanical check decides only what it can see. Rows marked M are decided by the checks below on every run. Rows marked A or H have no self-assessment on this site yet and are shown as not assessed; a tool does not decide them. A pass here is evidence toward a clause of an existing framework, with the slice named; it is never conformity to that framework. The checks cover output text. The decision layer (CC-6.6) is read from the weekly probe record (check decision.probe): fold rates are printed and, in record-only mode, never block. The probe measures a reference model (qwen3-coder-30b on Amazon Bedrock, bare and with a short stance instruction), not the model this assembly runs on; under the same design three larger models folded on 0 of 72 runs each (experiment 04, v5), so the fold rate is a property of the probed model and the evidence is the test and the record. Deploy gating: hugo.yml: the conformity checks run on the built output after the Hugo build and before the Pages artifact is uploaded; a failing check fails the build job and the previous build stays live (since the first push with this workflow, 2026-10-07).
| req | title | mark | state | evidence (this run) or self-assessment | maps to |
|---|---|---|---|---|---|
| CC-4.1 | Documented conformity testing programme | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 9(1)-(2); AIA Art 72(1); DORA Art 24(1) |
| CC-4.2 | Requirements under test with metric and threshold | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 9(8); NIST-AI-RMF MEASURE 1.1 |
| CC-4.3 | Named programme owner | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | NIST-AI-RMF GOVERN 2.1 |
| CC-5.1 | Tests run against the deployed assembly | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 15(1); DORA Art 24(2) |
| CC-5.2 | Component versions in every run record | M | pass | 20 pages and 8 component groups hashed; rule table matches the sha256 recorded in the site config | AIA Art 12(1); DORA Art 9(4)(e) |
| CC-5.3 | Blast-radius controls on the serving system | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | DORA Art 26(5) |
| CC-6.1 | Steady state measured before pressure | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | NIST-AI-RMF MEASURE 2.5 (nearest) |
| CC-6.2 | Pressure conditions from each relevant class | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 15(5); AIA Art 55(1)(b) (nearest) |
| CC-6.3 | Prediction record before each run | M | pass | this site publishes no prediction records | AIA Art 9(8) |
| CC-6.4 | Unannounced arms and invariance gap | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | DORA Art 26(1) (nearest) |
| CC-6.5 | Grader tested against known pass and fail cases | A | pass | 6 fixtures: 4 known-fail (one per rule), 2 known-pass; 0 failures | NIST-AI-RMF MEASURE 2.5 (nearest) |
| CC-6.6 | Decision-layer tests under scripted pressure, fold rate reported | A | partial | no decision-layer probe record yet. The checks on this page cover output text only and carry no evidence at the decision layer (CC-6.6) | AIA Art 15(1); AIA Art 15(5); AIA Art 9(8); DORA Art 26(2) (nearest) |
| CC-7.1 | Full test set on every change, before serving | M | pass | this run: trigger=push; deploy gated by the conformity job: True (since 2026-10-07); the deploy job runs only after this job passes | AIA Art 9(6)-(7); DORA Art 9(4)(e); DORA Art 25(1); NIST-CSF PR.PS-01 |
| CC-7.2 | Staged release with tests at each stage | H | pass | 0 placeholder patterns on 20 published files | DORA Art 9(4)(e) (nearest) |
| CC-7.3 | Full test set on a fixed interval | M | partial | no scheduled run yet; cron 37 6 * * 1 every 7 days is configured | DORA Art 24(6); GDPR Art 32(1)(d); AIA Art 72(2) |
| CC-7.4 | Regression set grows; removal documented | A | pass | 6 fixtures: 4 known-fail (one per rule), 2 known-pass; 0 failures | NIST-CSF ID.IM-03 |
| CC-7.5 | Published attack methods added within [60] days | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 55(1)(b) (nearest) |
| CC-8.1 | Knowledge items have an owner and a source of record | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 10(2) (nearest) |
| CC-8.2 | Last-verified date; stale items withdrawn | M | partial | 14 pages dated within 90 days; 2 older; 3 without a date. Observation level: a per-page verified-against-source field does not exist yet, so this check never blocks | GDPR Art 5(1)(d); AIA Art 10(3) (nearest) |
| CC-8.3 | Freshness interval per class of knowledge | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | GDPR Art 5(1)(d) (nearest) |
| CC-8.4 | Knowledge conformity tests graded against the source | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 15(1) (nearest) |
| CC-8.5 | Verification against the source, never a summary | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | GDPR Art 5(1)(d) (nearest) |
| CC-8.6 | Statements of fact traceable to a knowledge item | A | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 13(1) (nearest) |
| CC-9.1 | Pass or fail not decided by the producer alone | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 14(1); DORA Art 24(4) |
| CC-9.2 | Independent outside tester | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | DORA Art 26(8); DORA Art 27 |
| CC-10.1 | Nonconformities tracked; closure needs a passing rerun | A | pass | 0 open, 0 closed; closed without a passing rerun: 0 | NIST-CSF ID.IM-03; DORA Art 24(5) |
| CC-10.2 | Serious incidents passed to incident reporting | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | AIA Art 73(1); DORA Art 19(1) |
| CC-11.1 | Run record with the required fields | M | pass | all eight CC-11.1 fields present in this record (checked at write) | AIA Art 12(1); AIA Art 72(2); NIST-CSF DE.CM-09 |
| CC-11.2 | Run records kept | M | partial | CC-11.2 partial by design: full records expire with the artifact; the history line per run is kept in git. | AIA Art 19(1); AIA Art 18(1) |
| CC-11.3 | Published self-assessment labelled as such | H | pass | pages that present themselves as a self-assessment carry the label "Self-assessment, not a certification": 0 missing | AIA Art 43(2) (nearest) |
| CC-11.4 | Second rater invited; disagreements published | H | not assessed | no self-assessment exists for this site yet; a tool does not decide this row | DORA Art 24(4) (nearest) |
| CC-12.1 | Each requirement marked mechanical, assisted or manual | H | pass | 36 requirements marked; mechanical rows without a check id: none | AIA Art 14(1) |
| CC-12.2 | Mechanical requirements in machine-readable form | A | pass | 36 requirements marked; mechanical rows without a check id: none | NIST-CSF PR.PS-01 (nearest) |
| CC-12.3 | Output-boundary rule table with false-positive tests | M | pass | 0 blocking-tier hits over 20 published files; rules in blocking tier: service-closer, filler-idiom, hook-opener; 4 rules with a recorded false-positive test (site config); blocking rules without one: none | AIA Art 15(1); NIST-CSF PR.PS-01 |
| CC-12.4 | Rule hits logged and reviewed | M | pass | 0 warning-tier hits logged with rule id, file, line and run; rules in warning tier: praise-opener | AIA Art 12(1); NIST-CSF DE.CM-09 |
| CC-12.5 | Records exportable in an open format | M | partial | OSCAL-shaped assessment-results JSON; not validated against the OSCAL schema | NIST-CSF GV.OC (nearest) |
| check | result | detail and evidence |
|---|---|---|
| rules.blocking | pass | 0 blocking-tier hits over 20 published files; rules in blocking tier: service-closer, filler-idiom, hook-opener |
| rules.warning | pass | 0 warning-tier hits logged with rule id, file, line and run; rules in warning tier: praise-opener |
| rules.tests | pass | 4 rules with a recorded false-positive test (site config); blocking rules without one: none praise-opener: 4 hits, 4 legitimate, 2026-10-07, decision warn tier on the site tier until the regex is tightened to openers only service-closer: 0 hits, 0 legitimate, 2026-10-07, decision block filler-idiom: 0 hits, 0 legitimate, 2026-10-07, decision block hook-opener: 0 hits, 0 legitimate, 2026-10-07, decision block |
| placeholders | pass | 0 placeholder patterns on 20 published files |
| predictions.hashes | n/a | this site publishes no prediction records |
| site.consistency | pass | 19 pages checked; 0 failures, 6 observations blog/chaos-engineering-for-behaviour/index.html: not listed in llms.txt blog/knowledge-infrastructure/index.html: not listed in llms.txt blog/the-cheating-moved/index.html: not listed in llms.txt blog/the-stance-layer-is-still-toil/index.html: not listed in llms.txt book/index.html: not listed in llms.txt call/index.html: not listed in llms.txt |
| components | pass | 20 pages and 8 component groups hashed; rule table matches the sha256 recorded in the site config site commit e790b6993b46 rule table sha256 ee94ff54d716... (vestige-kit dc36b9c) requirements.json sha256 4f1704b2a0b0... |
| trigger.push | pass | this run: trigger=push; deploy gated by the conformity job: True (since 2026-10-07) hugo.yml: the conformity checks run on the built output after the Hugo build and before the Pages artifact is uploaded; a failing check fails the build job and the previous build stays live (since the first push with this workflow, 2026-10-07). |
| deploy.gated | pass | the deploy job runs only after this job passes hugo.yml: the conformity checks run on the built output after the Hugo build and before the Pages artifact is uploaded; a failing check fails the build job and the previous build stays live (since the first push with this workflow, 2026-10-07). |
| trigger.schedule | pending | no scheduled run yet; cron 37 6 * * 1 every 7 days is configured |
| page.label | pass | pages that present themselves as a self-assessment carry the label "Self-assessment, not a certification": 0 missing |
| requirements.marks | pass | 36 requirements marked; mechanical rows without a check id: none marks: M 10, A 14, H 12 |
| grader.fixtures | pass | 6 fixtures: 4 known-fail (one per rule), 2 known-pass; 0 failures |
| knowledge.freshness | partial | 14 pages dated within 90 days; 2 older; 3 without a date. Observation level: a per-page verified-against-source field does not exist yet, so this check never blocks blog/ai-means-nothing/index.html: newest date 2026-02-08, 243 days old (interval 90) blog/self-hosted-ai-pipeline-one-session/index.html: newest date 2026-02-08, 243 days old (interval 90) book/index.html: no date on the page call/index.html: no date on the page index.html: no date on the page |
| decision.probe | pending | no decision-layer probe record yet. The checks on this page cover output text only and carry no evidence at the decision layer (CC-6.6) |
| findings.closure | pass | 0 open, 0 closed; closed without a passing rerun: 0 |
| record.fields | pass | all eight CC-11.1 fields present in this record (checked at write) |
| record.retention | partial | CC-11.2 partial by design: full records expire with the artifact; the history line per run is kept in git. GitHub Actions artifacts, 90 days conformity/latest.json in git, indefinite |
| record.format | partial | OSCAL-shaped assessment-results JSON; not validated against the OSCAL schema |
A finding opens when a check fails and closes only when a later run passes that check without the hit (CC-10.1). Closed findings stay listed.
| id | status | check | where | requirements | opened / closed |
|---|---|---|---|---|---|
| None. | |||||
Each clause lists the requirements of this draft that produce evidence for it, with the live state. "Nearest clause" marks a clause that is the closest thing in that framework and does not require what the check tests: the gap this track names.
| clause | evidence from this tier |
|---|---|
| Art 9(1)-(2) | CC-4.1 not assessed risk management system as a continuous iterative process |
| Art 72(1) | CC-4.1 not assessed documented post-market monitoring system |
| Art 9(8) | CC-4.2 not assessed testing against prior defined metrics and probabilistic thresholds CC-6.3 pass prior defined metrics and thresholds CC-6.6 partial testing against prior defined metrics |
| Art 15(1) | CC-5.1 not assessed consistent performance of the system as placed on the market CC-6.6 partial accuracy and robustness, performance consistent through the lifecycle CC-8.4 not assessed nearest clause; it does not require this CC-12.3 pass consistent performance at the output |
| Art 12(1) | CC-5.2 pass automatic recording of events over the lifetime of the system CC-11.1 pass automatic recording of events CC-12.4 pass logging |
| Art 15(5) | CC-6.2 not assessed resilience against attempts to alter use, outputs or performance CC-6.6 partial resilience against attempts to alter outputs |
| Art 55(1)(b) | CC-6.2 not assessed nearest clause; it does not require this CC-7.5 not assessed nearest clause; it does not require this |
| Art 9(6)-(7) | CC-7.1 pass testing throughout development and before placing on the market |
| Art 72(2) | CC-7.3 partial evaluate continuous compliance throughout the lifetime CC-11.1 pass collection of data on performance throughout the lifetime |
| Art 10(2) | CC-8.1 not assessed nearest clause; it does not require this |
| Art 10(3) | CC-8.2 partial nearest clause; it does not require this |
| Art 13(1) | CC-8.6 not assessed nearest clause; it does not require this |
| Art 14(1) | CC-9.1 not assessed human oversight CC-12.1 pass which decisions a person takes |
| Art 73(1) | CC-10.2 not assessed reporting of serious incidents |
| Art 19(1) | CC-11.2 partial logs kept for at least six months |
| Art 18(1) | CC-11.2 partial documentation kept 10 years |
| Art 43(2) | CC-11.3 pass nearest clause; it does not require this |
| clause | evidence from this tier |
|---|---|
| Art 24(1) | CC-4.1 not assessed digital operational resilience testing programme |
| Art 24(2) | CC-5.1 not assessed testing of ICT systems supporting critical functions |
| Art 9(4)(e) | CC-5.2 pass documented ICT change management CC-7.1 pass ICT change management with testing before deployment CC-7.2 pass nearest clause; it does not require this |
| Art 26(5) | CC-5.3 not assessed risk management measures for threat-led tests on live production systems |
| Art 26(1) | CC-6.4 not assessed nearest clause; it does not require this |
| Art 26(2) | CC-6.6 partial nearest clause; it does not require this |
| Art 25(1) | CC-7.1 pass appropriate tests on ICT systems |
| Art 24(6) | CC-7.3 partial appropriate tests at least yearly |
| Art 24(4) | CC-9.1 not assessed tests by independent parties, internal or external CC-11.4 not assessed nearest clause; it does not require this |
| Art 26(8) | CC-9.2 not assessed external testers for threat-led tests |
| Art 27 | CC-9.2 not assessed requirements for testers |
| Art 24(5) | CC-10.1 pass remediation of issues identified in tests |
| Art 19(1) | CC-10.2 not assessed reporting of major ICT-related incidents |
| clause | evidence from this tier |
|---|---|
| MEASURE 1.1 | CC-4.2 not assessed approaches and metrics for measurement selected |
| GOVERN 2.1 | CC-4.3 not assessed roles and responsibilities documented |
| MEASURE 2.5 | CC-6.1 not assessed nearest clause; it does not require this CC-6.5 pass nearest clause; it does not require this |
| clause | evidence from this tier |
|---|---|
| PR.PS-01 | CC-7.1 pass configuration management CC-12.2 pass nearest clause; it does not require this CC-12.3 pass configuration management of the rule table |
| ID.IM-03 | CC-7.4 pass improvements from lessons learned CC-10.1 pass lessons learned |
| DE.CM-09 | CC-11.1 pass monitoring of software and services CC-12.4 pass monitoring |
| GV.OC | CC-12.5 partial nearest clause; it does not require this |
| clause | evidence from this tier |
|---|---|
| Art 32(1)(d) | CC-7.3 partial regular testing, assessing and evaluating |
| Art 5(1)(d) | CC-8.2 partial accuracy; kept up to date CC-8.3 not assessed nearest clause; it does not require this CC-8.5 not assessed nearest clause; it does not require this |
| run (UTC) | trigger | site commit | overall | open | warnings |
|---|---|---|---|---|---|
| 2026-10-09T10:09:41Z | push | e790b6993b46 | pass | 0 | 0 |
Full run records are workflow artifacts (90 days); the history line per run and the open findings live in latest.json in git. Records are OSCAL-shaped (assessment-results with observations and findings) and not yet validated against the OSCAL schema.
conformity: latest run 2026-10-09T10:09:41Z, 0 open, pass