Arc Skills / Configuration and change control
Check verification evidence and plan re-verification after a change
Arc Skills compares a proposed engineering change with pinned verification evidence, then plans the smallest credible re-verification. The output separates records that can be reused, records needing a delta argument, and obligations needing a targeted rerun.
Check existing evidence after a change
Use Arc Skills to identify evidence that may be stale after the supplied engineering change.
Inputs:
- Before/after requirement, component, interface, and software configurations
- Test definitions, Runs, reports, and article revisions pinned to each result
- Project evidence-applicability rules and the change rationale
Return an evidence register with the tested claim, changed mechanism, applicability verdict, and missing provenance.
Keep historical Runs intact. Explain why each record survives, needs analysis, or cannot support the target buildPlan targeted re-verification
Use Arc Skills to plan targeted re-verification for the supplied engineering change.
Inputs:
- Before/after change records and affected requirement baseline
- Existing Tests, Runs, evidence configurations, and applicability findings
- Target article, approved acceptance criteria, facilities, and owners
Return a change-to-requirement-to-evidence action table and a dependency-aware execution sequence.
Reuse unaffected results with a reason, seek delta analysis where sufficient, and rerun only unestablished claimsWhat you provide and what you get
| What you have | How it is used | What you get |
|---|---|---|
| Change C-7 and requirement baseline | Shows the actual threshold, actuator, and connector differences | Candidate affected obligations |
| Pinned Test and Run records | Connects each old result to a specific article and criterion | Evidence applicability register |
| Target article and change rationale | Tests whether changed mechanisms reach the measured claim | Reuse, delta analysis, or focused rerun plan |
One change package, three evidence treatments
Illustrative engineering example.
The illustrative C-7 package relaxes a valve timing limit from 2 s to 3 s, replaces actuator A1 with A2, and changes connector J1 pinout from revision A to B. Software parser 2.0 is unchanged. The target article contains A2 and J1 revision B.
R-4 before: Valve closes within 2 s; target revision: within 3 s.
E-1: valve closure 1.8 s on A1, old 2 s criterion.
E-2: valve closure 2.3 s on A1, old result Fail.
E-3: parser format unit test Pass on software 2.0.
E-4: environmental report on article with J1 revision A.
Target: A2, J1 revision B, parser 2.0.| Old evidence and tested configuration | Changed mechanism and applicability | Owner / smallest action |
|---|---|---|
| E-1: 1.8 s on A1, R-4 ≤2 s | Old measurement also meets ≤3 s numerically, but A2 changes closure mechanism; target applicability not established. | Valve owner: rerun closure timing on A2 under target criterion. |
| E-2: 2.3 s on A1, old Fail | May meet new ≤3 s threshold on old A1 if Run is valid; still cannot establish A2 timing. | Verification owner: retain failed history, review validity, then include A2 in targeted rerun. |
| E-3: parser Pass on software 2.0 | Parser and its format obligation are unchanged; independent of actuator and J1 electrical contacts on supplied design. | Software owner: document continuity and reuse E-3 for this tested claim. |
| E-4: environmental test with J1 rev A | J1 rev B may change installation or contact behavior; effect on tested environmental claim unknown. | Connector owner: compare mechanical/electrical delta; rerun affected exposure only if analysis cannot bridge it. |
The relaxed threshold does not erase E-2’s recorded failure under the old rule. A valid 2.3 s observation could support the new 3 s limit for A1, but the new A2 actuator breaks direct application to the target article. E-1 and E-2 therefore both lead to a focused A2 timing test, not a campaign-wide rerun.
E-3 is reusable only for the parser-format claim and only while its software revision and assumptions remain pinned. E-4 needs an owner’s connector delta assessment; a pinout change does not automatically invalidate every environmental result, but the old installation cannot be accepted by proximity alone.
Sequence: approve the target requirement and article baseline; review E-2 validity and the J1 delta; run A2 timing; run only any environmental condition the J1 analysis cannot cover. Interrupted or incomplete new Runs remain pending evidence.
Work from the change mechanism to the evidence
- Pin before/after requirement, component, interface, Test, Run, and article revisions.
- Identify which measured behaviors or assumptions each changed element can affect.
- Classify each old record as reusable, conditionally reusable through analysis, or unable to establish the target claim.
- Order narrow owner decisions and target tests; retain screened-out records with continuity reasons.
Questions about this task
Does a graph link make every connected Test stale?
No. The link finds candidates. The actual change mechanism and tested behavior decide applicability; E-3 is screened out for its unchanged parser claim.
Can a former failure become a pass under a relaxed limit?
The original result stays a failure under the old baseline. A valid raw measurement may support a new criterion for the tested build, but it does not prove the changed actuator.
Sources and further reading
- NASA Systems Engineering Handbook: Primary lifecycle and verification guidance; example change decisions are illustrative.
- ECSS-E-ST-10-02C verification: Official identification and scope; any project clause assessment needs its controlled edition and tailoring.