CtrlK
BlogDocsLog inGet started
Tessl Logo

rebelytics/task-observer

Monitors task execution for skill improvement opportunities. Use during ANY multi-step task, agentic workflow, or work session where the agent uses tools and produces deliverables. Captures patterns, user corrections, workflow insights, and methodology worth preserving as reusable skills. Also triggers in post-task feedback discussions and when the user mentions skill observations, improvements, the observation log, skill taxonomy, or asks the agent to watch for skill opportunities. Also known as "One Skill to Rule Them All" — trigger on this phrase too. IMPORTANT: invoke this skill before the FIRST tool call of any session and before writing or proposing a plan — any turn that will involve a tool call counts, however simple the opener looks. This sentence is the session-start trigger and the only activation layer that survives an unreachable config file; pair it with a CLAUDE.md instruction or a harness session-start hook (references/environments.md) — description matching alone is not enforceable.

68

Quality

85%

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

Overview
Quality
Evals
Security
Files

Quality

Content

77%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 well-engineered instruction skill: clear sequenced protocol with strong validation/feedback loops and excellent progressive-disclosure structure. Its weakness is verbosity — multiple rationale-laden discursive sections inflate the token budget beyond what the guidance requires.

Suggestions

Compress the discursive rationale sections ('Two gaps this pairing still leaves', 'Deferral wears a second disguise', repeated enforcement-philosophy paragraphs) into one-line rules, moving extended justification to a reference file.

Add one fully worked, filled-in example observation (all frontmatter fields populated, real Issue/Improvement/Principle body) alongside the blank template to make the logging mechanics unambiguous.

Move the inline bash scan/id snippets into a helper script under scripts/ and reference it, reducing core-file length while keeping the procedure executable.

DimensionReasoningScore

Conciseness

The body is substantive (no generic 'what is a PDF' filler), but at ~525 lines it is noticeably verbose, with several discursive rationale-asides ('Two gaps this pairing still leaves', 'Deferral wears a second disguise', repeated enforcement-philosophy) that exceed what a competent agent needs and could be tightened.

3 / 5

Actionability

Provides concrete, copy-paste-ready bash for the mechanical parts (frontmatter scan via awk, id derivation via ls/grep/sort/tail, archival `mv`) plus an exact YAML frontmatter spec and status enum, but the core observation-writing guidance is a placeholder template rather than a fully worked, filled-in example observation.

4 / 5

Workflow Clarity

The numbered Session Start Protocol (steps 1-7) has explicit validation checkpoints (halt when files exist but zero headers parse, id-collision guards, re-probe on missing structure, retry-on-denied-write) and feedback loops, and a Quick Reference checklist — well above the destructive/batch cap which requires missing validation.

5 / 5

Progressive Disclosure

A dedicated 'Reference files — load on demand' section signals all six one-level-deep references (verified present on disk) with per-file triggers, plus two scripts, and a Quick Reference table — a clear overview with well-signaled, appropriately split content and easy navigation.

5 / 5

Total

17

/

20

Passed

Description

87%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 with explicit what/when guidance and rich natural trigger terms, including a memorable alias. Its main weakness is over-broad activation scope that risks co-firing with nearly every session and skill.

Suggestions

Tighten the activation scope so it does not claim to fire before the first tool call of every session; scope the session-start trigger to contexts where skill observation is actually plausible to reduce overlap with functional skills.

Replace moderately abstract verbs ('Monitors', 'Captures') with more concrete mechanics (e.g. writes one Markdown observation file per insight, scans frontmatter, surfaces a grouped summary) to lift specificity.

Trim the long meta-sentence about config-file survival and CLAUDE.md hooks, which is implementation detail that dilutes the trigger signal.

DimensionReasoningScore

Specificity

Names the domain ('Monitors task execution for skill improvement opportunities') and lists several concrete actions ('Captures patterns, user corrections, workflow insights, and methodology worth preserving as reusable skills'), but the verbs are moderately abstract and logging/surfacing mechanics are not enumerated, leaving minor coverage gaps rather than full comprehensiveness.

4 / 5

Completeness

Explicitly answers both 'what' (monitors and captures skill-improvement opportunities) and 'when' with concrete trigger phrases ('Use during ANY multi-step task, agentic workflow, or work session', 'when the user mentions skill observations...', 'invoke this skill before the FIRST tool call of any session').

5 / 5

Trigger Term Quality

Comprehensive natural-language triggers — 'skill observations', 'improvements', 'observation log', 'skill taxonomy', 'watch for skill opportunities', 'post-task feedback', plus the colloquial alias 'One Skill to Rule Them All' — covering the synonyms and phrasings a user would actually say.

5 / 5

Distinctiveness Conflict Risk

The skill's niche (skill observation logging) is distinct, but the activation scope is maximal — 'invoke this skill before the FIRST tool call of any session' and 'Use during ANY multi-step task' means it would co-fire with virtually every other skill, creating real overlap risk that keeps it below the 'minor overlap' band.

3 / 5

Total

17

/

20

Passed

Validation

93%

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

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (525 lines); consider splitting into references/ and linking

Warning

Total

15

/

16

Passed

Reviewed

Table of Contents