Atri10/executor

skill:superpowers on steroids

3

stars

1

commits

Shell

primary language

Sep 3, 2026

updated

skill
skills
superpowers
Browse cluster: AI-Driven Code & Agent Development

README

The Executor

An initiative-scoped workflow system for coding agents: it takes a major idea from intake through architecture, spec, plan, execution, and review, with a strict per-initiative ID namespace, separated thinking and execution stores, and evidence-backed completion.

Nothing floats. Every document, task, review, and ruling carries an ID that names the initiative it belongs to. A body of work gets one Initiative; the initiative owns a folder, an ID namespace, and every document produced about it.

Works with any agent that can read skill files from a directory.

Install

Clone the repo and copy the skills/ directories into whatever directory your agent loads skills from:

git clone git@github.com:Atri10/executor.git
cp -R executor/skills/* <your-agents-skills-dir>/

Then invoke the root router:

/skill:executor

or just say "start an initiative" — normal requests route by the skill's frontmatter description.

Skills

SkillPhaseOutput
executorRouter + contractLoads the right phase skill, defines the ID namespace
executor-initiativeIntakeInitiative folder, charter
executor-discoveryDiscoveryResearch, options comparison
executor-architectureArchitecture, DesignArchitecture, ADRs, interfaces, component designs
executor-specSpecificationSpec, risks, verification strategy
executor-planningPlanningPlans with tasks
executor-executionExecutionCommits, reports, ledger
executor-reviewReviewVerdicts, findings, rulings
executor-verificationVerificationEvidence of working software
executor-handoffHandoffMerged branch, updated indexes

Phases compress, they never vanish. A small initiative can produce a charter and a spec in one exchange and skip discovery — but skipping is a stated decision recorded in the charter, not an omission.

The Two Stores

flowchart LR
    subgraph THINK["docs/executor/ - tracked"]
        C["Charter"] --> R["Research, Options"]
        R --> A["Architecture, ADRs, Interfaces"]
        A --> D["Design"]
        D --> S["Spec"]
        S --> P["Plans"]
    end
    subgraph EXEC[".executor/ - git-ignored by default"]
        L["Ledger, Rulings"]
        B["Briefs"]
        RP["Reports"]
        V["Diffs, Verdicts"]
    end
    P -->|"each task dispatch"| B
    B --> RP
    RP --> V
    V --> L
  • docs/executor/ — the thinking record. Git-tracked: charter, research, architecture, decisions, interfaces, design, spec, risks, plans. A reader who clones the repo gets the complete reasoning.
  • .executor/ — the execution record. Git-ignored by default, safe to commit if you choose: task briefs, implementer reports, review diffs and verdicts, the progress ledger, rulings. Never deleted — the reasoning is the point.

The split is durability-of-audience, not durability-of-value.

The ID Namespace

INIT-0004                      the initiative
INIT-0004-CHTR-01              its charter
INIT-0004-RSCH-02              a research note
INIT-0004-SPEC-01-R07          requirement 7 inside that spec
INIT-0004-P01                  a plan
INIT-0004-P01-T03              task 3 of that plan
INIT-0004-P01-T03-R02          review round 2 of that task

Addressable requirements are what let a review finding name the exact contract it violates, and what lets a plan task declare precisely which requirements it discharges.

Safety

.executor/ may be committed, so everything in both stores is written as though it will be public. Credentials, tokens, and personal data never go into any Executor artifact — a redacted existence statement and a safe path instead. skills/executor/references/safety.md defines the required scan before any handoff.

Project docs

DocPurpose
CONTRIBUTING.mdHow to change skills, references, and scripts
CODE_OF_CONDUCT.mdParticipation standards and enforcement
SECURITY.mdReporting vulnerabilities and how artifact secret-hygiene works
CHANGELOG.mdNotable changes, newest first

License

MIT

Contributors

Atri10

1 commits

Atri10/executor

skill:superpowers on steroids

3

stars

1

commits

Shell

primary language

Sep 3, 2026

updated

skill
skills
superpowers
Browse cluster: AI-Driven Code & Agent Development

README

The Executor

An initiative-scoped workflow system for coding agents: it takes a major idea from intake through architecture, spec, plan, execution, and review, with a strict per-initiative ID namespace, separated thinking and execution stores, and evidence-backed completion.

Nothing floats. Every document, task, review, and ruling carries an ID that names the initiative it belongs to. A body of work gets one Initiative; the initiative owns a folder, an ID namespace, and every document produced about it.

Works with any agent that can read skill files from a directory.

Install

Clone the repo and copy the skills/ directories into whatever directory your agent loads skills from:

git clone git@github.com:Atri10/executor.git
cp -R executor/skills/* <your-agents-skills-dir>/

Then invoke the root router:

/skill:executor

or just say "start an initiative" — normal requests route by the skill's frontmatter description.

Skills

SkillPhaseOutput
executorRouter + contractLoads the right phase skill, defines the ID namespace
executor-initiativeIntakeInitiative folder, charter
executor-discoveryDiscoveryResearch, options comparison
executor-architectureArchitecture, DesignArchitecture, ADRs, interfaces, component designs
executor-specSpecificationSpec, risks, verification strategy
executor-planningPlanningPlans with tasks
executor-executionExecutionCommits, reports, ledger
executor-reviewReviewVerdicts, findings, rulings
executor-verificationVerificationEvidence of working software
executor-handoffHandoffMerged branch, updated indexes

Phases compress, they never vanish. A small initiative can produce a charter and a spec in one exchange and skip discovery — but skipping is a stated decision recorded in the charter, not an omission.

The Two Stores

flowchart LR
    subgraph THINK["docs/executor/ - tracked"]
        C["Charter"] --> R["Research, Options"]
        R --> A["Architecture, ADRs, Interfaces"]
        A --> D["Design"]
        D --> S["Spec"]
        S --> P["Plans"]
    end
    subgraph EXEC[".executor/ - git-ignored by default"]
        L["Ledger, Rulings"]
        B["Briefs"]
        RP["Reports"]
        V["Diffs, Verdicts"]
    end
    P -->|"each task dispatch"| B
    B --> RP
    RP --> V
    V --> L
  • docs/executor/ — the thinking record. Git-tracked: charter, research, architecture, decisions, interfaces, design, spec, risks, plans. A reader who clones the repo gets the complete reasoning.
  • .executor/ — the execution record. Git-ignored by default, safe to commit if you choose: task briefs, implementer reports, review diffs and verdicts, the progress ledger, rulings. Never deleted — the reasoning is the point.

The split is durability-of-audience, not durability-of-value.

The ID Namespace

INIT-0004                      the initiative
INIT-0004-CHTR-01              its charter
INIT-0004-RSCH-02              a research note
INIT-0004-SPEC-01-R07          requirement 7 inside that spec
INIT-0004-P01                  a plan
INIT-0004-P01-T03              task 3 of that plan
INIT-0004-P01-T03-R02          review round 2 of that task

Addressable requirements are what let a review finding name the exact contract it violates, and what lets a plan task declare precisely which requirements it discharges.

Safety

.executor/ may be committed, so everything in both stores is written as though it will be public. Credentials, tokens, and personal data never go into any Executor artifact — a redacted existence statement and a safe path instead. skills/executor/references/safety.md defines the required scan before any handoff.

Project docs

DocPurpose
CONTRIBUTING.mdHow to change skills, references, and scripts
CODE_OF_CONDUCT.mdParticipation standards and enforcement
SECURITY.mdReporting vulnerabilities and how artifact secret-hygiene works
CHANGELOG.mdNotable changes, newest first

License

MIT

Contributors

Atri10

1 commits

Languages

Shell

50.1%

JavaScript

39.4%

HTML

10.6%