ECSS GSE Acceptance (space-systems/ecss/q20-gse-acceptance)
Use when the task is the clause 5.8.4.2 acceptance of ground support equipment
in ECSS-Q-ST-20C: the acceptance tests and acceptance reviews have been run
against the agreed GSE requirements, and the equipment is either released for
use, released with what it cannot do written down, or held.
Domain quick reference
- The verdict on a result and the measurement behind it are two facts, and
they disagree more often than anyone expects. A line reading pass against a
reading that misses its bound is the finding the review exists to catch.
- A margin is worth more than a verdict. Two requirements both pass; one
passes by a factor and one by a whisker, and only the second tells the
operations team what to watch.
- A value landing exactly on its bound is inside it. The bound is the bound,
and the last few bits of a sensor conversion are a representation question,
not an engineering one.
- Criticality decides what a method may be. Equipment that holds, lifts or
powers flight hardware is shown behaving; analysis and review of design show
intent, and intent does not carry a load.
- Coverage is over the requirement set, not over the results. Twelve results
against four requirements leaves whatever was never exercised unexercised,
and a not-run result on a requirement reopens it whatever ran earlier.
- A waiver does not turn a failure into a pass. It turns it into a limitation
the equipment carries, and the release says so on its face.
Workflow
- Validate the agreed requirement set: unique identifiers, a criticality and a
planned method from the closed vocabularies, and a bound with the side of it
the equipment must stay on wherever the requirement is quantitative.
- Validate the acceptance results the same way, refusing a result against a
requirement nobody agreed and an outcome outside the vocabulary.
- Grade each quantitative result on its own measurement, compute the margin,
and raise the contradiction where a reported pass does not survive it.
- Raise a failure with no approved waiver behind it, and a method the result
used where the requirement planned another.
- Grade method adequacy against criticality separately, since a safety-critical
requirement closed on paper is a finding whatever the paper concluded.
- Compute acceptance coverage, letting a not-run result reopen a requirement,
and compare it with the required value through a named tolerance.
- Sort waivered failures into limitations, then return released,
released-with-limitations or not-released.
Pitfalls
- Taking the outcome column at face value. The measurement is the evidence and
the outcome is somebody's reading of it; the two are compared, not conflated.
- Recording a pass and discarding the margin. The equipment that passed by
0.01 is the equipment that will fail next season, and nobody kept the number.
- Counting results instead of requirements. Three runs against one requirement
leaves the other three untouched, and a result count hides that.
- Letting an analysis close a lifting or pressure requirement. Those are shown
under load or under pressure, and no calculation substitutes for the sight.
- Reading a waiver as a pass. The equipment is limited, the limitation travels
with it, and the release states it rather than burying it in the minutes.
- Accepting a coverage figure that just misses its target as close enough. The
comparison is made through a stated tolerance or not at all.
Behavior contract (gate 3)
The requirement and result validation, measurement grading with margins,
pass-versus-measurement contradiction, unwaivered failures, method
substitution, criticality-based method adequacy, acceptance coverage,
limitation extraction and the release verdict are exercised by the gate 3
contract test: scripts/test_q20_gse_acceptance.py against
scripts/q20_gse_acceptance_logic.py (stdlib unittest, offline). Run:
python3 scripts/test_q20_gse_acceptance.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and
paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
1---2name: q20-gse-acceptance3description: "Evaluate the acceptance of ground support equipment under ECSS-Q-ST-20C clause 5.8.4.2: grade each acceptance result against the agreed GSE requirement it answers, keep the measured margin rather than only the verdict, refuse a reported pass whose own measurement misses the bound, insist that a safety-critical requirement be shown by test or inspection instead of argued on paper, compute acceptance coverage over the requirement set, and separate a failure carried on an approved waiver from one that stops the equipment. Use when GSE has finished its acceptance campaign and someone has to say whether it may be used. Trigger: ecss, q-st-20c-clause-5-8-4-2, gse-acceptance-test-grading, gse-acceptance-margin, gse-acceptance-coverage, gse-waivered-limitation, gse-release-for-use."4license: Apache-2.05compliance: STANDARDS-REF6standards:7 - id: ecss8 reference-only: true9gated: false10domain: space-systems11pack: space-systems12compatibility: "agentskills.io SKILL.md; any SKILL.md host (Claude Code, Hermes, OpenClaw)"13metadata:14 source: https://skillmd.com/skills/ashfordeou/q20-gse-acceptance15 domain: space-systems16 subdomain: ecss17 tags: [ecss, q-st-20c-quality-assurance-scope, q20-gse-acceptance, gse-acceptance-test-grading, gse-acceptance-margin, gse-acceptance-coverage, gse-waivered-limitation, gse-release-for-use]18 version: 0.1.019 author: Aero Agent Skills20---2122# ECSS GSE Acceptance (space-systems/ecss/q20-gse-acceptance)2324Use when the task is the clause 5.8.4.2 acceptance of ground support equipment25in ECSS-Q-ST-20C: the acceptance tests and acceptance reviews have been run26against the agreed GSE requirements, and the equipment is either released for27use, released with what it cannot do written down, or held.2829## Domain quick reference3031- The verdict on a result and the measurement behind it are two facts, and32 they disagree more often than anyone expects. A line reading pass against a33 reading that misses its bound is the finding the review exists to catch.34- A margin is worth more than a verdict. Two requirements both pass; one35 passes by a factor and one by a whisker, and only the second tells the36 operations team what to watch.37- A value landing exactly on its bound is inside it. The bound is the bound,38 and the last few bits of a sensor conversion are a representation question,39 not an engineering one.40- Criticality decides what a method may be. Equipment that holds, lifts or41 powers flight hardware is shown behaving; analysis and review of design show42 intent, and intent does not carry a load.43- Coverage is over the requirement set, not over the results. Twelve results44 against four requirements leaves whatever was never exercised unexercised,45 and a not-run result on a requirement reopens it whatever ran earlier.46- A waiver does not turn a failure into a pass. It turns it into a limitation47 the equipment carries, and the release says so on its face.4849## Workflow50511. Validate the agreed requirement set: unique identifiers, a criticality and a52 planned method from the closed vocabularies, and a bound with the side of it53 the equipment must stay on wherever the requirement is quantitative.542. Validate the acceptance results the same way, refusing a result against a55 requirement nobody agreed and an outcome outside the vocabulary.563. Grade each quantitative result on its own measurement, compute the margin,57 and raise the contradiction where a reported pass does not survive it.584. Raise a failure with no approved waiver behind it, and a method the result59 used where the requirement planned another.605. Grade method adequacy against criticality separately, since a safety-critical61 requirement closed on paper is a finding whatever the paper concluded.626. Compute acceptance coverage, letting a not-run result reopen a requirement,63 and compare it with the required value through a named tolerance.647. Sort waivered failures into limitations, then return released,65 released-with-limitations or not-released.6667## Pitfalls6869- Taking the outcome column at face value. The measurement is the evidence and70 the outcome is somebody's reading of it; the two are compared, not conflated.71- Recording a pass and discarding the margin. The equipment that passed by72 0.01 is the equipment that will fail next season, and nobody kept the number.73- Counting results instead of requirements. Three runs against one requirement74 leaves the other three untouched, and a result count hides that.75- Letting an analysis close a lifting or pressure requirement. Those are shown76 under load or under pressure, and no calculation substitutes for the sight.77- Reading a waiver as a pass. The equipment is limited, the limitation travels78 with it, and the release states it rather than burying it in the minutes.79- Accepting a coverage figure that just misses its target as close enough. The80 comparison is made through a stated tolerance or not at all.8182## Behavior contract (gate 3)8384The requirement and result validation, measurement grading with margins,85pass-versus-measurement contradiction, unwaivered failures, method86substitution, criticality-based method adequacy, acceptance coverage,87limitation extraction and the release verdict are exercised by the gate 388contract test: scripts/test_q20_gse_acceptance.py against89scripts/q20_gse_acceptance_logic.py (stdlib unittest, offline). Run:90python3 scripts/test_q20_gse_acceptance.py9192## Compliance9394- ECSS standards are freely downloadable (ESA); cite the source and95 paraphrase per standards-map.yaml.96- compliance: STANDARDS-REF, gated: false.