Aoa Session Progression Lift

aoa-session-progression-lift

Intent

Use this skill to author a PROGRESSION_DELTA from reviewed evidence.

This is not score inflation. It is an evidence-backed overlay that says what mastery axes moved, what should hold or reanchor, what baseline the movement is compared against, and what small unlock hints are now safer.

Trigger boundary

Use this skill when:

• a reviewed session generated meaningful mastery evidence
• the route needs progression legibility without mutating source role profiles
• donor harvest, checkpoint closeout, or a reviewed handoff produced provisional

axis hints that must be accepted, rejected, or carried without becoming growth claims by themselves

• movement needs to be distinguished from first observation, stale baseline,

missing baseline, or regression against a known prior state

• quest or RPG reflection would help continuation
• the output needs to stay small, evidence-backed, and multi-axis

Do not use this skill when:

• there is no reviewed evidence
• the only input is a checkpoint axis hint, mood note, or generated summary

that has not been reread against reviewed session evidence

• the route wants comparative movement while refusing to name a baseline or

admit that the baseline is missing, stale, or contested

• the request wants one global power number
• progression is being used as hidden routing policy
• the route is trying to mint authority rights without evidence

Inputs

• reviewed session artifact or harvest packet
• named evidence refs
• optional donor harvest output, checkpoint-closeout stage result, or

provisional axis hints

• current-session boundary for separating stale or neighboring evidence
• baseline ref, prior delta ref, or an explicit no-baseline marker
• relevant role or cohort context if known
• existing progression baseline if available

Outputs

• PROGRESSION_DELTA with axis movement, verdict, and optional unlock hints
• baseline posture such as baseline_ref, prior_delta_ref,

first_observed, baseline_missing, baseline_stale, or baseline_contested

• one axis table where each meaningful axis names movement, evidence refs,

evidence posture, and any defer or no-movement reason

• optional automation-readiness hint when reviewed evidence supports it
• optional rank reflection note if evidence is strong enough
• quest hooks or chronicle stub when useful
• negative or cautionary evidence when a hold, reanchor, or downgrade is more honest than advance
• one PROGRESSION_DELTA_RECEIPT using references/stats-event-envelope.md

and references/progression-delta-receipt-schema.yaml

• one CORE_SKILL_APPLICATION_RECEIPT using

references/core-skill-application-receipt-schema.yaml

Procedure

1. collect reviewed evidence refs
2. separate confirmed current-session evidence from checkpoint hints,

closeout-handoff cues, generated summaries, stale residue, and neighboring session evidence

1. declare the movement basis before assigning any axis movement: baseline ref,

prior delta ref, first-observed posture, missing baseline, stale baseline, or contested baseline

1. assess movement qualitatively across boundary_integrity,

execution_reliability, change_legibility, review_sharpness, proof_discipline, provenance_hygiene, and deep_readiness

1. mark each touched axis with evidence posture: confirmed, contested,

provisional, stale, not_current_session, or no_movement

1. emit a verdict: advance, hold, reanchor, or downgrade
2. when the baseline is missing, stale, or contested, avoid comparative advance

language unless a specific current-session axis has confirmed evidence and the limitation remains visible

1. when evidence is only provisional or contested, prefer hold, reanchor, or an

explicit defer reason over invented advance

1. name small unlock hints only when evidence supports them
2. keep any automation-readiness hint small, descriptive, and non-authoritative
3. allow negative, zero, and cautionary movement
4. map ability or feat hints only as reflection, not as ownership transfer
5. hand off to aoa-quest-harvest only when a repeated reviewed quest unit

survives as a final promotion question; otherwise keep quest hooks as reflection or defer them

1. emit one PROGRESSION_DELTA_RECEIPT when the delta closes, keeping the

receipt descriptive, evidence-linked, and smaller than the progression packet

1. when the finish path closes, emit one CORE_SKILL_APPLICATION_RECEIPT

that points back to the bounded progression receipt and records one finished kernel-skill application

Contracts

• progression remains evidence-backed
• multi-axis only; no authoritative universal score
• movement claims must name their baseline posture; missing baseline can support

first-observed evidence notes, not fake comparative growth

• checkpoint, donor, and closeout hints may focus the review but do not become

progression claims until reviewed evidence supports them

• rank labels are descriptive, not sovereign
• unlock hints must stay reviewable and small
• progression does not replace owner-layer truth or routing authority
• progression does not greenlight automation by itself
• progression receipts stay descriptive and append-only
• generic core receipts stay subordinate to the progression receipt and do not

become progression authority

• receipt corrections use supersedes rather than silent overwrite

Risks and anti-patterns

• inventing progress from mood
• promoting checkpoint axis hints or generated summaries into movement without

reviewed evidence

• calling a first observation an improvement without naming the absent baseline
• treating absence of a problem as evidence of mastery movement
• using progression as policy
• granting authority without cited evidence
• flattening multi-axis growth into one number
• confusing quest flavor with durable proof
• treating progression hints as if they were approval or schedule rights
• treating a progression receipt as if receipt emission itself proved growth

Verification

• confirm all meaningful axis claims cite reviewed evidence
• confirm the delta declares baseline posture before comparative movement claims
• confirm provisional, stale, contested, and no-movement axes are marked rather

than smoothed into progress

• confirm the verdict matches the evidence
• confirm zero or negative movement remains allowed
• confirm unlock hints are small and explicit
• confirm no universal score is introduced
• confirm any emitted receipt stays multi-axis, evidence-linked, and non-sovereign
• confirm any emitted CORE_SKILL_APPLICATION_RECEIPT points to the

progression detail receipt and stays finish-only

Technique traceability

Manifest-backed techniques:

• AOA-T-0084 from 8Dionysus/aoa-techniques at 3b1d5d623569aa4920b87280d0db0e911d2e29d5 using path techniques/continuity/donor-harvest/progression-evidence-lift/TECHNIQUE.md and sections: Intent, Inputs, Outputs, Contracts, Validation
• AOA-T-0085 from 8Dionysus/aoa-techniques at 3b1d5d623569aa4920b87280d0db0e911d2e29d5 using path techniques/continuity/donor-harvest/multi-axis-quest-overlay/TECHNIQUE.md and sections: Outputs, Risks, Validation

Adaptation points

Project overlays may add:

• local axis notes
• role-affinity hints
• local unlock classes
• local axis posture vocabulary, provided it still distinguishes confirmed

reviewed evidence from provisional or stale hints

• local baseline vocabulary, provided it still distinguishes a reviewed prior

baseline from first observation, missing baseline, stale baseline, or contested baseline

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: 8Dionysus
• Source: 8Dionysus/aoa-skills
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.