aspiers/ai-config / .agents/skills/comparing-open-source-projects
comparing-open-source-projects
UnverifiedResearches and compares free and open source software projects in a supplied domain, using either user-named candidates or a discovered shortlist. Use when evaluating FOSS alternatives, comparing GitHub popularity, maintenance, licences, domain fit, trade-offs, or producing a linked HTML comparison report.
Comparing Open Source Projects
Compare projects against the user's actual domain and use case, not against a context-free feature checklist. Treat current popularity, maintenance, and licensing claims as evidence that must be checked rather than remembered.
Establish the comparison set
- When the user names exact candidates, compare those candidates. Do not add or replace projects unless the user also asks for discovery.
- When the user supplies only a domain, define a transparent, representative shortlist from current research. Search across project websites, canonical forges, ecosystem directories, and credible roundups; then verify every candidate against primary sources.
- State the inclusion criteria and important exclusions. Do not imply that a shortlist is exhaustive.
- Ask one focused clarification when different interpretations of the domain would materially change the candidates or recommendation. Otherwise state reasonable assumptions in the report.
- If a named candidate is source-available, proprietary, abandoned, or not a real match, retain it and explain the limitation rather than silently substituting another project.
Collect comparable evidence
Use the same observation date and activity window for every project. Prefer canonical project and repository sources; use secondary sources to discover claims, not to establish them.
For every candidate, establish:
- Identity and fit: canonical project and repository, relevant edition or component, supported use cases, deployment model, and important scope boundaries.
- GitHub popularity: the star count of the canonical primary repository,
with its repository link and an as-of UTC date. Do not sum an organisation,
plugins, forks, or mirrors. If GitHub is not canonical, report
N/Aand identify any GitHub mirror separately. Stars measure attention, not quality. - Maintenance: archived or deprecation state, date of the latest default- branch commit, latest stable release, commit/release activity over a common window, and evidence of issue or pull-request responsiveness where useful. Distinguish active development, maintenance-only maturity, sporadic work, inactivity, and insufficient evidence. Do not infer maintenance from stars or from one recent automated commit.
- Licence: exact licence name and SPDX identifier when one exists, linked to the governing licence text. Check multiple licences, exceptions, component-specific terms, and community-versus-enterprise boundaries. GitHub's detected licence is a discovery hint, not a substitute for reading the repository's licence files. Clearly distinguish OSI/FSF-recognised open source licences from source-available terms and avoid presenting the report as legal advice.
- Decision factors: capabilities relevant to the requested domain, maturity, integrations, operational burden, documentation/community, major limitations, and the kinds of users for whom the project is or is not a good fit.
Link factual claims to evidence. Use exact dates and counts where available, and mark unknowns instead of estimating. Explain the basis of qualitative maintenance ratings so a mature low-churn project is not automatically treated as abandoned.
Analyse and recommend
Normalise terminology and compare like with like. Separate observations from judgement, identify material evidence gaps, and explain trade-offs instead of manufacturing a universal winner. Base recommendations on the user's stated needs and call out when licensing or maintenance makes an otherwise strong candidate unsuitable.
Use matrix-comparison for the ranked HTML matrix, separate capability columns, accessible traffic-light ratings and multiple required-feature filters. It owns shared presentation and controls. Keep software-specific evidence and judgements in this skill. Leave non-ordinal facts uncoloured unless an explicit suitability judgement applies.
Produce the report
Read the FOSS reporting requirements before writing. They add the summary fields and licence/maintenance evidence to the shared matrix contract. Use the matrix-comparison renderer and template; this skill's legacy template path points to that same shared asset.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |