Ejecuta cualquier Skill en Manus
con un clic

Ejecuta cualquier Skill en Manus con un clic

e2sa-add-skill

Estrellas0

Forks0

Actualizado9 de julio de 2026 a las 21:10

Scaffold and register a new E2SA skill so the steps are never half-done. Use when the human or agent wants to create a new capability — "add a skill", "create a new skill", "scaffold a skill", "make this reusable as a skill", "distill X into a skill". Creates the SKILL.md with correct frontmatter, registers it in the AGENTS.md capability catalog (the manual registry), seeds a Changelog, and stops for human review before commit. Human-gated.

Instalación

Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.

Ejecutar en Manus

Fuente

jingtao-lbl

jingtao-lbl/E2SA-SPADE

Abrir repositorio de GitHub Ver repositorios del creador

Descarga

Ejecutar en Manus

Ocupaciones relacionadasSOC

Basado en la clasificación ocupacional SOC

Otras ocupaciones informáticasOcupaciones informáticas y matemáticas·SOC 15-1299

SKILL.md

readonly

name	e2sa-add-skill
description	Scaffold and register a new E2SA skill so the steps are never half-done. Use when the human or agent wants to create a new capability — "add a skill", "create a new skill", "scaffold a skill", "make this reusable as a skill", "distill X into a skill". Creates the SKILL.md with correct frontmatter, registers it in the AGENTS.md capability catalog (the manual registry), seeds a Changelog, and stops for human review before commit. Human-gated.
allowed-tools	["Read","Glob","Grep","Write","Edit","Bash"]

e2sa-add-skill

Scaffolds a new skill AND registers it, because the two-step "create the SKILL.md and update the AGENTS.md catalog" is easy to half-do (it was, for e2sa-add-data-source). Design: docs/design/09_skill_evolution.md.

When to fire

The human asks to add/create/scaffold a skill.
The agent recognizes a recurring procedure worth making reusable (propose it; do not self-create without surfacing).

Procedure

Name + scope (check the existing names FIRST). Before choosing <name>, list the existing skills (ls .claude/skills/ + skim the AGENTS.md catalog) and avoid a name that collides with or is confusingly similar to one already there. A shared prefix + verb invites mis-firing and tab-completion confusion (e.g. e2sa-validate vs e2sa-validate-adapter — renamed to e2sa-audit-adapter for exactly this reason). Pick a name whose distinct verb/object makes the boundary obvious. If a near-neighbor still exists after naming, give both descriptions an explicit NOT for … — use <other> boundary line so the harness's trigger matcher can't confuse them. Kebab-case; framework-general → e2sa-<name>, project-specific → scope under the project. The directory name becomes the /<name> slash command; confirm with the human if unsure.
Write .claude/skills/<name>/SKILL.md with the frontmatter:
- name: matching the directory.
- description: — this IS the trigger the harness matches; write it as "what it does + when to use it + example phrasings." Highest-leverage line; get it right.
- allowed-tools: — the minimal tool set.
- Body: a short purpose line, "When to fire", a numbered "Procedure", "Guardrails", and a ## Changelog.
Register it (the manual step): add a row to the "Capability catalog" table in AGENTS.md (| | invoke when… | notes |). This is the human-facing registry; auto-discovery handles the functional side.
Seed the ## Changelog with a dated "Initial version" line, and (optional) a signals: note pointing to where evidence for future refinement will live (memory/knowledge/lessons/).
Verify the registry (mechanical backstop): run python tools/check_skill_registry.py — it must exit 0. This confirms the new skill has its AGENTS.md catalog row, its frontmatter name: matches the directory, and it carries a ## Changelog. If it reports DRIFT, you skipped step 3; fix and re-run.
Stop for human review, then commit (.claude/skills/<name>/ + AGENTS.md) with a plain message. Do not commit unreviewed.

Guardrails

Register or it drifts: never finish without the AGENTS.md catalog row. tools/check_skill_registry.py (step 5) is the enforceable backstop — wire it into CI / pre-commit so the invariant holds even when this skill isn't used.
Distinct names: check existing skill names before adding (step 1); a near-duplicate name (shared prefix + verb) makes the harness and humans pick the wrong skill. Disambiguate by renaming and/or NOT for … boundary lines in both descriptions.
Minimal procedure: a skill encodes the easy-to-get-wrong conventions, not an essay. If it is one obvious step, it may not need to be a skill.
Framework skills stay project-agnostic (no project sites/paths baked in); project-specific logic lives under projects/<project>/.
Mark non-user-invocable / auto-fire skills explicitly in the description (like e2sa-lessons-capture).
Human-gated: propose + review before commit.

Changelog

2026-06-17: Initial version (per docs/design/09_skill_evolution.md).
2026-06-17: Broadened the trigger to include "distill X into a skill" (a natural team/09 phrasing for creating a new skill from existing knowledge).
2026-06-17: Added step 5 — mechanical registry verification via tools/check_skill_registry.py (drift backstop; the motivating e2sa-add-data-source gap is now enforceable, not just convention).
2026-06-23: Step 1 now requires checking existing skill names first to avoid confusingly-similar names (the e2sa-validate vs e2sa-validate-adapter collision, renamed to e2sa-audit-adapter); added the "distinct names" guardrail and the NOT for … boundary-line remedy.

Más de este repositorio

mismo repositorio

e2sa-refine-skill

jingtao-lbl/E2SA-SPADE

Improve an existing E2SA skill from accumulated evidence, human-gated. Use when a skill missed a step or needs updating, when a lesson/feedback in a skill's domain has piled up, or during a periodic skill review — "refine the X skill", "improve this skill", "the X skill should have caught Y", "review the skills". Gathers the signal (lessons, run-journal outcomes, feedback), proposes a concrete SKILL.md diff, and STOPS for human approval; it never self-applies. After approval it edits, appends a Changelog line, and commits.

2026-07-090

e2sa-add-adapter

jingtao-lbl/E2SA-SPADE

Implement a new E2SA data-source adapter (the code that fetches a source and parses it into Observation records). Use when writing the adapter for a source that has already been onboarded/inspected — "build the adapter for X", "write the X adapter", "implement the fetch/parse for X", or the coding step after e2sa-add-data-source. Walks the BaseAdapter contract (list_available / fetch / parse_to_schema), the FetchResult single-file-vs-whole-package shape, the Observation + Provenance schema, canonical units + depth + serves, ADAPTER_REGISTRY registration, the fixture + test, and the parse gotchas already learned (sentinels, CRS tiers, obs_id uniqueness, schema drift, idempotency).

2026-07-080

e2sa-add-connector

jingtao-lbl/E2SA-SPADE

Implement a new E2SA data-center connector (the per-data-center access layer that owns auth + search + whole-package fetch under Option C). Use when wrapping a data center not yet connected — "add a connector for X", "wrap the PANGAEA/Earthdata/Zenodo API", or when migrating an existing self-fetching adapter onto a connector. Walks the BaseConnector contract, the probe-the-API-first step, the two search patterns (server-side filter vs enrich + bbox coverage-ratio), the two fetch shapes (BagIt-zip vs file-by-file), the reusable conventions (urllib + User-Agent, raw layout, token-free reads, on-disk fast-path, manual fallback), CONNECTOR_REGISTRY registration, and the mocked-urlopen + E2E_LIVE test split. Distilled from the arctic_data_center and ess_dive connectors.

2026-07-080

e2sa-add-data-source

jingtao-lbl/E2SA-SPADE

Add a new data source (data center, archive, or dataset) into an E2SA project. Use when the human or agent has found a dataset to bring in, is inspecting a new archive, or says "add a data source", "add this dataset", "I found data at X", or after downloading data from a center not yet documented. Walks the full procedure: inventory check, download whole package, inspect (format/metadata-standard/CRS/license/citation), write the project-named source card, update the sources index and data-center registry, capture a finding, and credential setup. Always uses the latest dataset version and records a full citation.

2026-07-080

e2sa-audit-adapter

jingtao-lbl/E2SA-SPADE

Audit an E2SA adapter's data output against the REAL source and propose source-grounded corrections (applied only on human approval). Use after building or changing an adapter, after a parse, when staged data looks off, or for a periodic adapter review — "audit/validate the X adapter", "check the data X produces", "is above_stdm correct", "review the adapters for data-quality". Runs the e2sa/qc checks (serves-subset, value-range, subsurface-depth, citation-not-synthesized, self-describing) on a real parse, surfaces the value distributions, cross-checks variables/units/citation against the source, and PROPOSES a source-grounded fix for each finding (never invents) that is applied only on human approval (propose-first, like e2sa-refine-skill). NOT for run-level shippability (run skeleton / notebooks / secrets / a run_id) — that is `e2sa-validate`. Distilled from the 2026-06-23 above_stdm fabrications (reflection 20260623s).

2026-07-080

e2sa-file-task

jingtao-lbl/E2SA-SPADE

File a data-preparation task into projects/<project>/tasks/ when the current work needs site data that isn't already prepared. Use whenever the agent (not just the human) recognizes a need for forcing data, surface dataset assembly, observations, or QC/analysis that warrants its own request with provenance. Always sets status=requested and surfaces to the human; never auto-executes the filed task in the same fire.

2026-07-080

name	e2sa-add-skill
description	Scaffold and register a new E2SA skill so the steps are never half-done. Use when the human or agent wants to create a new capability — "add a skill", "create a new skill", "scaffold a skill", "make this reusable as a skill", "distill X into a skill". Creates the SKILL.md with correct frontmatter, registers it in the AGENTS.md capability catalog (the manual registry), seeds a Changelog, and stops for human review before commit. Human-gated.
allowed-tools	["Read","Glob","Grep","Write","Edit","Bash"]

e2sa-add-skill

When to fire

The human asks to add/create/scaffold a skill.
The agent recognizes a recurring procedure worth making reusable (propose it; do not self-create without surfacing).

Procedure

Name + scope (check the existing names FIRST). Before choosing <name>, list the existing skills (ls .claude/skills/ + skim the AGENTS.md catalog) and avoid a name that collides with or is confusingly similar to one already there. A shared prefix + verb invites mis-firing and tab-completion confusion (e.g. e2sa-validate vs e2sa-validate-adapter — renamed to e2sa-audit-adapter for exactly this reason). Pick a name whose distinct verb/object makes the boundary obvious. If a near-neighbor still exists after naming, give both descriptions an explicit NOT for … — use <other> boundary line so the harness's trigger matcher can't confuse them. Kebab-case; framework-general → e2sa-<name>, project-specific → scope under the project. The directory name becomes the /<name> slash command; confirm with the human if unsure.
Write .claude/skills/<name>/SKILL.md with the frontmatter:
- name: matching the directory.
- description: — this IS the trigger the harness matches; write it as "what it does + when to use it + example phrasings." Highest-leverage line; get it right.
- allowed-tools: — the minimal tool set.
- Body: a short purpose line, "When to fire", a numbered "Procedure", "Guardrails", and a ## Changelog.
Register it (the manual step): add a row to the "Capability catalog" table in AGENTS.md (| | invoke when… | notes |). This is the human-facing registry; auto-discovery handles the functional side.
Seed the ## Changelog with a dated "Initial version" line, and (optional) a signals: note pointing to where evidence for future refinement will live (memory/knowledge/lessons/).
Verify the registry (mechanical backstop): run python tools/check_skill_registry.py — it must exit 0. This confirms the new skill has its AGENTS.md catalog row, its frontmatter name: matches the directory, and it carries a ## Changelog. If it reports DRIFT, you skipped step 3; fix and re-run.
Stop for human review, then commit (.claude/skills/<name>/ + AGENTS.md) with a plain message. Do not commit unreviewed.

Guardrails

Register or it drifts: never finish without the AGENTS.md catalog row. tools/check_skill_registry.py (step 5) is the enforceable backstop — wire it into CI / pre-commit so the invariant holds even when this skill isn't used.
Distinct names: check existing skill names before adding (step 1); a near-duplicate name (shared prefix + verb) makes the harness and humans pick the wrong skill. Disambiguate by renaming and/or NOT for … boundary lines in both descriptions.
Minimal procedure: a skill encodes the easy-to-get-wrong conventions, not an essay. If it is one obvious step, it may not need to be a skill.
Framework skills stay project-agnostic (no project sites/paths baked in); project-specific logic lives under projects/<project>/.
Mark non-user-invocable / auto-fire skills explicitly in the description (like e2sa-lessons-capture).
Human-gated: propose + review before commit.

Changelog

2026-06-17: Initial version (per docs/design/09_skill_evolution.md).
2026-06-17: Broadened the trigger to include "distill X into a skill" (a natural team/09 phrasing for creating a new skill from existing knowledge).
2026-06-17: Added step 5 — mechanical registry verification via tools/check_skill_registry.py (drift backstop; the motivating e2sa-add-data-source gap is now enforceable, not just convention).
2026-06-23: Step 1 now requires checking existing skill names first to avoid confusingly-similar names (the e2sa-validate vs e2sa-validate-adapter collision, renamed to e2sa-audit-adapter); added the "distinct names" guardrail and the NOT for … boundary-line remedy.