CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

antithesis-debug

Run a bash script inside an Antithesis run at a chosen moment with `snouty runs exec`, from the terminal and without a browser. Use it for scripted forensics on a run: process and container inspection, on-disk evidence, gdb attach, and general debugging. Covers command usage, the live-session requirement, common workflows and errors, and how virtual time and perturbation limit what a debugger can see.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

88%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.

A high-quality operational skill: copy-paste-ready commands, an explicitly sequenced workflow with validation checkpoints and feedback loops, and a well-organized one-level-deep reference bundle. The main improvements are removing cross-section repetition, isolating version-sensitive details, and fixing the dangling references/properties.md pointer.

Suggestions

Consolidate the repeated 404/"Moment not warm" retry and first-exec --source-run-id guidance into one place (e.g. the Rewarm progress or Common Errors section) and cross-reference it from the Example Workflow and Known issues, to trim duplicated tokens in conciseness.

Fix the dangling reference: references/moments.md points to references/properties.md, which does not exist in the bundle — either add the file or drop the pointer, and verify the 'antithesis-triage' mention resolves for a reader of the skill.

Move time-sensitive details (snouty 0.8.0 minimum, release 64.0, 'default is 10m as of release 64.0') into the frontmatter compatibility field or a dedicated version-notes line so the body stays stable as versions change.

DimensionReasoningScore

Conciseness

The body is dense with operational detail Claude would not know (snouty flags, rewarm behavior, jq pipelines) and avoids explaining known concepts, but guidance is repeated across sections — 404/"Moment not warm" retry advice appears in "Get a live session first", the Example Workflow, and "Known issues", and first-exec `--source-run-id` advice appears three times. Version-sensitive details ("The default is 10m (as of release 64.0)", snouty 0.8.0) are inline rather than isolated, and a few grammar slips ("Preference using a currently running run", "for your users request") suggest it could be tightened.

4 / 5

Actionability

Fully executable throughout: exact commands with realistic values (`snouty runs exec "$RUN_ID" -1520695250096255483 67.7622960046865 'uptime'`), ready-to-run jq filters, heredoc script files, tar+base64 file extraction, and worked kubectl/gdb examples in the references. The guidance covers the common cases (live run, past run, Kubernetes, cold moment).

5 / 5

Workflow Clarity

The Example Workflow gives a clearly sequenced 1→2a/2b→3→4→5 path with explicit validation checkpoints: checking `in_progress:ready` status, `runs wait --until ready`, retry-on-404 feedback loops, checking rewarm progress before retrying, and searching for output before re-running state-changing scripts. The Self-Review checklist before reporting and the mandatory user confirmation before `runs cancel` round out validation for the destructive cleanup step.

5 / 5

Progressive Disclosure

Good structure: the body is an operational overview that cleanly offloads depth to five real, one-level-deep, clearly signaled reference files (moments, gdb, kubernetes, virtual-time, errors), all of which exist. One dangling pointer — `references/moments.md` cites `references/properties.md`, which is not in the bundle — and the main file's length leave minor organization gaps.

4 / 5

Total

18

/

20

Passed

Description

92%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.

A strong description: it states exactly what the skill does, names the tool and concrete operations, and includes an explicit use-when clause with specific scenarios. Only minor room remains in broadening natural trigger synonyms.

DimensionReasoningScore

Specificity

The description lists multiple concrete actions — "Run a bash script inside an Antithesis run at a chosen moment", "process and container inspection, on-disk evidence, gdb attach, and general debugging" — giving comprehensive, domain-specific coverage rather than generic capability claims. The final sentence ("Covers command usage, the live-session requirement...") is slightly meta but does not dilute the concrete action list.

5 / 5

Completeness

Both halves are explicit: the "what" ("Run a bash script inside an Antithesis run at a chosen moment with `snouty runs exec`, from the terminal and without a browser") and a clear "when" clause ("Use it for scripted forensics on a run: process and container inspection, on-disk evidence, gdb attach, and general debugging") with concrete trigger scenarios.

5 / 5

Trigger Term Quality

Good natural-term coverage: "debugging", "forensics", "gdb", "process and container inspection", "Antithesis run", "bash script" are phrases users would plausibly say. A few natural synonyms like "troubleshoot" or "investigate a failing property" are missing, keeping it just below the comprehensive anchor.

4 / 5

Distinctiveness Conflict Risk

The description is anchored to a distinct niche — executing scripts inside Antithesis runs via `snouty runs exec` — so it would not plausibly trigger for other debugging or container skills. Antithesis-specific terms ("run", "moment", "virtual time", "perturbation") further isolate it.

5 / 5

Total

19

/

20

Passed

Validation

100%

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

antithesishq/antithesis-skills

Path

antithesis-debug/SKILL.md

Reviewed

Table of Contents

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.