All skills
simota avatar

/atlas

@965f4f9
by shingo imotasimota/agent-skills85 stars
15

Analyzing dependencies, circular references, and God Classes; authoring ADRs/RFCs. Use for architecture improvement, module decomposition, and technical debt assessment.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/atlas

This session only. Nothing lands on disk.

referencetechnical-debt-scoring.md

≈630 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Technical Debt Scoring

Load for a debt inventory, repayment plan or ROI comparison. The SKILL core contract owns TDR thresholds, the ≥15% capacity rule above 5% TDR, Cost-of-Delay ordering, and comprehension_debt; do not substitute a vendor's survey or a fixed example sprint budget for those rules.

Severity Matrix

Impact Low effort Medium effort High effort
High P0: Fix now P1: Next sprint P2: Plan
Medium P1: Next sprint P2: Plan P3: Backlog
Low P2: Plan P3: Backlog P4: Accept

Item Scoring

Priority = Impact × (6 - Cost), each scored 1–5. Support Impact with touchpoints, frequency of developer friction and related defects; support Cost with remediation effort, regression risk and prerequisite changes. Record production-incident risk (1–5) and the estimate's unit separately. Do not silently equate this ranking with TDR or the severity matrix.

Categories: Design, Code, Architecture, Test, Documentation, Infrastructure, Dependency. Use the affected quality/owner to choose; a security issue is not automatically medium merely because it is dependency debt.

Inventory / Repayment Contract

Inventory: total/high-priority counts and estimated effort; each TD-ID's category, location, Impact, Cost, Priority, proposed fix, dependencies and owner. Include comprehension_debt HIGH/MEDIUM/LOW with actual AI-authorship/review-depth evidence; its remediation is documentation/ADR backfill/review, not automatically refactoring.

Repayment plan: protected capacity and accountable owner; measurable goals; staged work linked to TD-IDs and dependencies; milestones; before/after success metrics. Avoid a standalone rewrite or an unfunded “20% rule.” Validate outcomes with defects, tests, cycle time and business impact rather than optimizing a single tool score.

ROI

  • Monthly avoidable cost: defect-fix time + extra development time + onboarding overhead, each × the applicable rate.
  • One-time fix cost: development + testing + review/deployment effort, each × rate.
  • Break-even months: fix cost / monthly savings.
  • 12-month ROI: (monthly savings * 12 - fix cost) / fix cost * 100.

Declare assumptions/ranges, currency and time window. If savings are non-positive, there is no positive break-even; if fix cost is zero, percentage ROI is undefined. Conclude Fix / Defer / Accept with the rationale. Industry-wide spending figures are not evidence of this repository's costs.

Source: SKILL.md on GitHub

No alerts13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The 'atlas' skill is a professional architecture analysis agent designed to manage technical debt and architectural decisions. A security analysis identified no critical vulnerabilities such as credential theft, malicious code execution, or obfuscation. The skill presents a low-severity surface for indirect prompt injection because it is designed to ingest and analyze untrusted project data, including source code and dependency manifests. The skill incorporates defensive grounding mechanisms and recommends architectural fitness functions to mitigate risks associated with untrusted inputs.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    1/13 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 965f4f9. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/atlas