agentsociety-create-env-module

Use when creating or revising a custom environment module, when an experiment needs an environment class that does not yet exist in the workspace, or when the module design must fit a simulation budget.

Quality

58%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./extension/skills/agentsociety-create-env-module/v1.0.0/SKILL.md

Quality

Content

63%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body delivers a clearly sequenced, validated workflow with concrete commands and a useful mistake/fix table, and it mostly trusts the reader's competence. Its weaknesses are the triple-stated no-inheritance rule and a reference graph where most linked files are absent from the bundle, undermining navigation.

Suggestions

Consolidate the no-inheritance rule into a single section (plus the pitfalls.md P5 pointer) and reduce the other two mentions to one-line cross-references, cutting roughly 15 lines of repetition.

Ship the referenced bundle files (stages/*.md, checklists/compatibility.md, artifacts/schema.md, subagent-prompts/planner.md, implementer.md, reviewer.md) or inline their critical content so the body's navigation links resolve.

Include one minimal copy-paste example of an @tool method with the required dict return shape and a sample create-env-module-validate invocation with real flag values.

Dimension	Reasoning	Score
Conciseness	The body is mostly dense and actionable, but the no-inheritance rule is stated in full three times — in Runtime Contract ("Inherit ONLY from EnvBase. Never subclass..."), in its own section ("the class must inherit directly from EnvBase"), and again in the Common Mistakes table — before also pointing to references/pitfalls.md P5. This repetition is more than the 'minor instances' of anchor 4, fitting anchor 3's 'could be tightened'.	3 / 5
Actionability	Concrete guidance is strong: exact output path ("custom/envs/<module>.py"), a full validation command with flags ("create-env-module-validate (flags: --file, --workspace, --class-name, --run-id, --json, --no-refresh-metadata)"), a step-by-step delegation procedure, and a mistake/fix table. It falls short of anchor 5 because commands use placeholders ($PYTHON_PATH, "...") and no copy-paste example of a @tool method or DesignSpec appears inline.	4 / 5
Workflow Clarity	The six-stage flow is clearly sequenced (intake → clarify → design → generate → validate → archive) with an explicit validation mandate ("Always run .agentsociety/bin/ags.py create-env-module-validate before finishing") and an error-recovery loop ("fix any remaining issues from the reviewer report"). It is not anchor 5 because per-stage checkpoints and failure-mapping details are deferred to stages/*.md rather than stated, leaving minor validation gaps in the body itself.	4 / 5
Progressive Disclosure	Section structure and signposting are decent (Stage Notes, Shared References with bolded pointers), but most referenced paths do not exist in the bundle — all five stages/.md, checklists/compatibility.md, artifacts/schema.md, and all three subagent-prompts/.md are missing, leaving navigation broken for the majority of links. Combined with pitfalls content duplicated inline in the Common Mistakes table, this fits anchor 3 (references present but the organization has real gaps) rather than anchor 4's 'minor organization gaps'.	3 / 5
	Total	14 / 20 Passed

Description

53%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description has excellent, explicit trigger guidance but omits any standalone statement of what the skill actually does — its capabilities are only implied by the first trigger clause. It is distinct within its niche yet lacks the natural term variations and concrete action list of top-scoring descriptions.

Suggestions

Add an explicit 'what' statement before the triggers, e.g. "Creates or repairs a single-file EnvBase environment module under custom/envs with @tool-decorated methods, then validates it with the create-env-module-validate CLI."

Include natural synonyms and example phrasings users would say, such as "simulation environment", "custom env", or example asks like a social-media or voting environment.

Mention the validation step in the description so the skill's end-to-end scope (generate then validate) is discoverable from the frontmatter alone.

Dimension	Reasoning	Score
Specificity	"creating or revising a custom environment module" names the domain and one concrete action, but the description stops there — sub-capabilities the body clearly supports (tool generation, validation via CLI, single-file output) are absent. It matches anchor 3 (domain + 1-2 concrete actions) rather than 4, which requires several listed actions.	3 / 5
Completeness	The "when" is explicit and well-structured (three "Use when..." clauses), but there is no standalone "what" statement — the capability is only weakly implied through the first trigger clause. This mirrors anchor 3's imbalance (one half explicit, the other weakly implied); it is not anchor 2 because a what is embedded, and not anchor 4 because that what is never explicitly stated.	3 / 5
Trigger Term Quality	Phrases like "environment module", "environment class", "simulation budget", and "experiment" are relevant, but common variations a user would naturally say — "simulation environment", "custom env", "new env class", or example domains like "social media module" — are missing. Coverage is partial, matching anchor 3 rather than the good coverage of anchor 4.	3 / 5
Distinctiveness Conflict Risk	"custom environment module ... that does not yet exist in the workspace" carves a distinct niche with minimal conflict risk, but "environment module" and "experiment" could overlap with closely related sibling skills (e.g. scan-modules, experiment-config), matching anchor 4 rather than the fully distinct anchor 5.	4 / 5
	Total	13 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: tsinghua-fib-lab/AgentSociety
Path: extension/skills/agentsociety-create-env-module/v1.0.0/SKILL.md
Commit: 8cc5bb9

Reviewed: 6 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.