AIM — PRESS KIT
AIM 4.7.7 · 29 October 2026.

POSITIONING
Turn your intent into a complete product.

ONE SENTENCE
AIM completes the agreed product with Ship by default, offers Loop for recurring owner feedback, and makes the work, dependencies and evidence visible in AIM UI.

SHORT DESCRIPTION
Agile Iteration Method (AIM) is an open-source method and tooling for AI-assisted software development. Describe the result you want. Ship takes the agreed product through planning, implementation, independent review and real verification before final acceptance. Choose Loop when recurring owner feedback should shape the next useful Done Increment. AIM UI is the local control room for Stories, the selected delivery, dependencies and evidence. Graph engineering connects responsibilities, prerequisites and proof. AIM UI Graph shows Story dependencies, evidence and external inputs; /aim graph provides the selected recorded execution view.

CURRENT CAPABILITIES
- Discover tools when needed, use supported native processing for large MCP/tool responses, and retain sources, counterevidence, partial errors and approvals. Small calls stay direct; bounded ordinary calls provide the fallback. Actual token and cost savings depend on the task and host.
- Keep unfinished Stories visible across deliveries, link corrections to original requirements, and return the verified original scope to its actual owner decision.
- Reassess user value and cumulative effort before diagnosis expands. Graph requires a current bounded assessment after repeated failed attempts; ongoing work also keeps the coordinator responsible.
- Keep product fixes separate from unknown method causes in read-only learning triage, preserving prior corrections and evidence boundaries.
- Open AIM UI with /aim ui to inspect User Stories, the selected delivery, Graph, review evidence and recorded progress alongside your coding agent. The UI remains Beta; copying a handoff does not start agents or grant approval.
- Ship is the default: complete the agreed product before final acceptance. Choose Loop for useful Done Increments with recurring owner feedback.
- Use Graph engineering to connect ownership, dependencies, delivery and independent proof. An agent owner can give feedback and decide in Strict Loop only under an explicit human mandate and its permitted delivery decisions.
- Use /aim graph to inspect delivery activities or native checkpoints, current proof, blockers and bounded assignments. AIM UI Graph shows Story dependencies and external inputs. Changed activity inputs identify work and evidence that need correction.
- Agree quality goals before implementation and test the first working core before building around it. A separate reviewer challenges a material core with an independent example.
- Verify the real user journeys through the delivered application. Keep raw observations behind the assessment, including complete-result time when it matters.
- Maintain a map of important user journeys and a runnable verification route so later changes can reuse the right checks.
- Reuse repository knowledge, compact confirmed checkpoints and still-applicable results. Existing approvals and answers stay bound to their scope.
- Recall relevant shared lessons during planning, review and diagnosis. Share verified reusable corrections with existing proof during ordinary delivery.
- Select Stories across Epics, review a coherent increment and keep user acceptance explicit. Closing an Epic remains a separate decision.

EXAMPLE
A customer needs reliable booking confirmation. A support specialist needs a useful failure reference. Those Stories can belong to different Epics and be delivered together. The plan includes delayed responses, repeated requests and the complete confirmation journey, with retained results for review.

INSTALLATION
Run in the project's terminal:
  npx skills add joneri/agile-iteration-method --skill agile-iteration-method
Choose the agent and installation scope, reload the session if needed, load the installed AIM skill and start with /aim start "<desired outcome>". AIM conditionally calibrates missing or stale repository knowledge and configures applicable project specialists and available role skills within that same start. Existing storage restrictions and overrides remain binding; empty projects start provisionally. Use /aim calibrate-repo or /aim configure-agents for later refreshes. The same instructions can be given to a coding agent. The portable skill is complete AIM; the reviewed adaptive installer remains available for broader repository setup.

FACTS
Name: Agile Iteration Method (AIM)
Release version: 4.7.7
Runtime contract: 2.0; backlog, state and installer schemas: 1.0
Platforms: Codex, GitHub Copilot and Claude Code
AIM UI: Beta; local control room
Creator: Jonas Eriksson
Source: https://github.com/joneri/agile-iteration-method
Website: https://joneri.github.io/agile-iteration-method/
Skill: https://skills.sh/joneri/agile-iteration-method/agile-iteration-method
Public releases: https://github.com/joneri/agile-iteration-method/releases

EVIDENCE AND AVAILABILITY
The retained Enigma M3 results compare GPT-6 Luna with AIM, scoring 96.5/100 in the latest assessment, with GPT-6 Luna without AIM at 49.8/100. Both used high reasoning effort, the same brief and the same scoring weights. The earlier AIM result was 93.5/100. The Luna-only run is from 30 September 2026; the latest assessed AIM delivery is from 18 October. Delivery conditions differ: this is a historical comparison, not a fresh paired experiment or proof of general superiority or lower development cost. The latest version of AIM includes later changes outside this application score; the assessed native delivery remained unfinished. Exact frozen versions, scope and sanitized evidence: https://github.com/joneri/agile-iteration-method/blob/main/docs/features/aim-enigma-evaluation.md. Original with-and-without-AIM comparison: https://github.com/joneri/agile-iteration-method/blob/main/docs/features/aim-3.2-enigma-m3-case.md.

The earlier AIM 3.2 paired study retains its original test conditions and evidence. AIM 4.7.7 is available as the current ordinary release. Install the complete skill with the command above or view the release at https://github.com/joneri/agile-iteration-method/releases/tag/v4.7.7.

LOGO
The supplied aim-4-logo.png is the established AIM campaign lockup: the golden/blue folded A identity, navy AIM lettering and a blue/cyan ribbon numeral 4 on a transparent background. Use on white or pale backgrounds. Preserve proportions, colours and the full lockup. Leave at least the height of the AIM capital letters as clear space. Check legibility at the intended size; do not use the entire wordmark as a tiny favicon. The versionless canonical repository artwork remains separate.
Generation: built-in image_gen, 12 October 2026. Full original prompt in logo-generation.txt. Raster PNG, 2172 × 724 pixels, RGBA; this package does not include a vector master.

PACKAGE CONTENTS
 aim-4-logo.png — transparent campaign logo
 aim-4-press-kit.txt — current positioning, facts and usage notes
 launch-copy-en.txt — English announcement draft
 launch-copy-sv.txt — Swedish announcement draft
 logo-generation.txt — historical generation method and exact prompt
 LICENSE-DOCS.txt — repository documentation licence

ATTRIBUTION
Agile Iteration Method (AIM), by Jonas Eriksson.
